moderated mediation hypothesis example

• Chapter 1: Introduction
• Chapter 2: Indexing
• Chapter 3: Loops & Logicals
• Chapter 4: Apply Family
• Chapter 5: Plyr Package
• Chapter 6: Vectorizing
• Chapter 7: Sample & Replicate
• Chapter 8: Melting & Casting
• Chapter 9: Tidyr Package
• Chapter 10: GGPlot1: Basics
• Chapter 11: GGPlot2: Bars & Boxes
• Chapter 12: Linear & Multiple
• Chapter 13: Ploting Interactions
• Chapter 14: Moderation/Mediation
• Chapter 15: Moderated-Mediation
• Chapter 16: MultiLevel Models
• Chapter 17: Mixed Models
• Chapter 18: Mixed Assumptions Testing
• Chapter 19: Logistic & Poisson
• Chapter 20: Between-Subjects
• Chapter 21: Within- & Mixed-Subjects
• Chapter 22: Correlations
• Chapter 23: ARIMA
• Chapter 24: Decision Trees
• Chapter 25: Signal Detection
• Chapter 26: Intro to Shiny
• Chapter 27: ANOVA Variance
• Download Rmd

Chapter 15: Moderated Mediation

Anthony n. washburn, 1 quick review of moderation and mediation, 1.1 moderation.

Basic Moderation Model

• Moderation tests the influence of a third variable (Z) on the relationship between X to Y
• X -> Y (depending on Z)
• For a review see Chapter 14: Mediation and Moderation

1.2 Mediation

Basic Mediation Model

• Mediation tests a hypothetical causal chain where the effect of one variable (X) on another variable (Y) is mediated, or explained, by a third variable (M)
• X -> M -> Y

2 What is moderated mediation?

2.1 conceptual definition.

Basic Moderated Mediation Model

• Moderated mediation tests the influence of a fourth (or more) variable on the mediated relationship between X and Y
• The effect of the mediator is moderated by another variable
• X -> M -> Y (depending on Z)
• The moderation can occur on any and all paths in the mediation model (e.g., a path, b path, c path, or any combination of the three)

2.2 Practical definition and example

• The more time one spends in graduate school, the more job offers they have when they graduate
• This relationship is explained by increased publications (i.e., the more time spent in grad school, the more publications one has, and the more publications one has, the more job offers they get)
• However, this causal chain may only work for people who spend their time in graduate school wisely (i.e., spend time with Professor Demos)
• How does spending time with Professor Demos impact the causal chain between time spent in graduate school, publications, and job offers? Let’s find out…

3 Moderated mediation data example

3.1 describe the dataset.

We are going to simulate a dataset that measured the following:

• X = Time spent in graduate school (we will change the name to “time” when we create the data frame)
• Z = Time spent (hours per week) with Professor Demos in class or in office hours
• M = Number of publications in grad school
• Y = Number of job offers

Our Example Moderated Mediation Model

3.2 Create the dataset

We are intentionally creating a moderated mediation effect here and we do so below by setting the relationships (the paths) between our causal chain variables and setting the relationships for our interaction terms

setwd("path of working directory here")

Here we are creating the values of our variables for each subject

Now we put it all together and make our data frame

3.3 Examine the dataset and prepare for regression analyses

install.packages("psych") #install this package if not already installed

Because we have interaction terms in our regression analyses, we need to mean center our IV and Moderator (Z)

4 Moderated mediation analyses using “mediation” package

We will first create two regression models, one looking at the effect of our IVs (time spent in grad school, time spent with Alex, and their interaction) on our mediator (number of publications), and one looking at the effect of our IVs and mediator on our DV (number of job offers).

Next, we will examine the influence of our moderating variable (time spent with Alex) on the mediation effect of time spent in grad school on number of job offers, through number of publications. To do this, we will examine the mediation effect for those who spend a lot of time with Alex versus those who spend little time with Alex.

4.1 Create the necessary regression models

We need two regression models to use the mediation package

One model specifies the effect of our IV (time spent in grad school) on our Mediator (number of publications) [and in our case, our moderator (time spent with Alex) and the interaction]

The other model specifies the effect of the IV (time spent in grad school) and Mediator (number of publications) (and possibly moderator as well) on our DV (number of job offers)

install.packages("mediation") #install this first if not already installed

• Significant main effect of time spent in grad school on number of publications
• Significant main effect of time spent with Alex on number of publications
• Significant interaction between time spent in grad school and time spent with Alex on number of publications
• Significant main effect of time spent in grad school on number of job offers
• No effect of time spent with Alex on number of job offers
• Significant main effect of number of publications on number of job offers
• No interaction between time spent in grad school and time spent with Alex on number of job offers

4.2 Examine the effect of our moderator on the mediation effect

In this mediation package we list the moderator as a covariate and set the levels to what we want

We can use the +/- 1SD from the mean (or another value that is theoretically important)

This allows us to view impact of the moderator on the direct and indirect effects

Lets look at grad students who spend little time with Alex first

For a review on bootstrapping techniques, see Efron, 2003

• ACME: Average Causal Mediation Effect [total effect - direct effect]
• ADE: Average Direct Effect [total effect - indirect effect]
• Total Effect: Direct (ADE) + Indirect (ACME)
• Prop. Mediated: Conceptually ACME / Total effect (This tells us how much of the total effect our indirect effect is “explaining”)

• Significant direct effect of time spent in grad school on job offers (for those who don’t spend a lot of time with Alex)
• Significant indirect effect of time spent in grad school on job offers through publications (for those who don’t spend a lot of time with Alex)

Now let’s look at grad students who spend a lot of time with Alex

• Significant direct effect of time spent in grad school on job offers (for those who spend a lot of time with Alex)
• Significant indirect effect of time spent in grad school on job offers through publications (for those who spend a lot of time with Alex)
• The indirect effect looks larger for those who spend a lot of time with Alex compared to those who don’t, but we can test this to make sure

The following code tests whether the difference between indirect effects at each level of the moderator is significantly different from zero

• We can see that the indirect effects are significantly different such that the effect of spending time in graduate school on getting job offers through publications is stronger for those students who spend a lot of time with Alex compared to those who do not
• There is no different in the size of the direct effects, however

4.3 Strengths and limitations of “mediation” package

• Code is fairly straightforward and makes intuitive sense in how to specify levels of moderators
• Compatible with many types of regression, including linear, glm, ordered, censored, quantile, GAM, and survival
• Limited in the types of moderated mediation models it can estimate
• Must include moderator in both models (meaning that you cannot model two of the most popular moderated mediation models, Hayes’ Model 7 and Model 14)
• Cannot handle highly complex mediational models with several causally dependent mediators and moderators
• However, structural equation model (SEM) programs can model more complex models, which we turn to next

5 Moderated mediation analyses using “lavaan” package

In “lavaan” we specify all regressions and relationships between our variables in one object

We can specify the effects we want to see in our output (e.g., direct, indirect, etc.)

We can also compute means and standard deviations for use in simple slopes analyses

After specifying all the necessary components, we fit the model using an SEM function

install.packages("lavaan") #install this first if not already installed

Now we take the specified models and all of the effects we want to estimate and run them through the SEM function. The SEM function allows a completely user-defined model to be fit to the data, like our specifically defined moderated mediation model (the SEM function was designed to fit structural equation models, but can also fit “regular” regression models as well).

• The first chunk of the output show fit indices related to SEM (not really applicable for our purposes)
• The second part of the output shows our regression formulas
• The end of the output shows the specified direct, indirect, total, proportion mediated effects

We can also call for bootstrapped confidence interval parameter estimates of all of our effects

• Our estimates and confidence intervals are almost identical to the “mediation” package estimates
• The difference is most likely a result of bootstrap estimation differences (e.g., lavaan uses bias-corrected but not accelerated bootstrapping for their confidence intervals)

5.1 Strengths and limitations of “lavaan” package

• Extremely customizable
• Can also model latent variables if your measurement model requires it
• Tedious! It took me several hours to figure out how the naming conventions worked
• A lot of up front coding required meaning you kind of need to know exactly what you’re looking for in your model

6 References and Links

6.1 references.

Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach . New York: The Guilford Press.

Michalak, N. (2016, July 29). Reproducing Hayes’ PROCESS models’ results in R . Retrieved from https://nickmichalak.blogspot.com/2016/07/reproducing-hayess-process-models.html

Rosseel, Y. (2017, February 24). Package ‘lavaan’ . Retrieved from https://cran.r-project.org/web/packages/lavaan/lavaan.pdf

Sales, A. C. (2017). Review: Mediation package in R. Journal of Educational and Behavioral Statistics, 42 , 1, 69-84.

Tingley, D., Yamamoto, T., Hirose, K., Keele, L., & Imai, K. (2014). Mediation: R package for causal mediation analysis .

6.2 Helpful Links

The Lavaan Package Website

R Markdown Cheatsheet

R Markdown Gallery

Shannon Library closed 5/11 - 5/13

Shannon Library will be closed to the public to allow for facilities work on Saturday, Sunday, and Monday. Check full hours and other locations here.

Getting Started with Moderated Mediation

In a previous post we demonstrated how to perform a basic mediation analysis. In this post we look at performing a moderated mediation analysis. The basic idea is that a mediator may depend on another variable called a "moderator". For example, in our mediation analysis post we hypothesized that self-esteem was a mediator of student grades on the effect of student happiness. We illustrate this below with a path diagram. We see a direct effect of grades on happiness, but also an indirect effect of grades on happiness through self-esteem . A mediation analysis helps us investigate and quantify that indirect effect.

But what if we suspect that, say, gender moderates that indirect effect? In other words, what if we think that the mediation effect of self-esteem might differ between females and males? To analyze that question we use moderated mediation . The difference between mediation and moderated mediation is that we include an interaction for the moderator in our models. Let's demonstrate using R . First we read in the data from our mediation analysis post, but this time with a gender variable added. Notice we format gender as a factor. This is required for the mediation code to work. (Note: this data and example are fake and just for illustration.)

Next we load the mediation package . If you don't already have the mediation package, run the install.packages function below. Otherwise you can skip it.

Now we define our mediator and outcome models with an interaction term for gender. The interaction needs to happen with both "treatment" and mediating variables. In this case, grades is our "treatment" and self-esteem is the mediator.

Notice this is just like the code in the mediation analysis post except we've added an interaction for gender in both models. The formula notation grades*gender is a short cut for writing grades + gender + grades:gender , where ":" is an interaction operator in R's formula syntax. An interaction allows the effect of grades and self-esteem to vary according to gender. Now we run our mediation as before using the mediate() function with 1000 simulations.

Finally we perform the moderated mediation using the test.modmed() function. This is where we perform the simulation draws to calculate uncertainty. The first argument is the output of the mediation analysis. The second and third arguments are the different levels of the moderators. Notice they each need to be a list object. The last argument specifies the number of simulations, where once again we set it to 1000. Technically we don't need to include this argument. By default the test.modmed() function will use the number of simulations specified in the original mediate() call.

Since we're using simulation to estimate uncertainty, your answer will differ slightly from the output above. The first section is a test of difference between the average causal mediation effects (ACME), i.e., the indirect effect of grades through self-esteem on happiness. The estimated difference is about -0.056, but the 95% confidence interval spans from -0.396 to 0.259. The difference is small and we don't have enough evidence to conclusively determine whether it's positive or negative. The second section is a test of difference between the average direct effects (ADE), i.e., the direct effect of grades on happiness. As with the indirect effect, we don't have enough evidence to conclude if the difference in direct effects between genders is positive or negative. In this case our moderator was a categorical variable but a moderator can also be continuous. We just have to specify different values of the moderator in the covariates arguments of test.modmed() . See the documentation of test.modmed() for an example by entering ?test.modmed in your R console.

• MacKinnon, D. (2008). Introduction to Statistical Mediation Analysis . Lawrence Erlbaum.
• R Core Team (2018). R: A Language and Environment for Statistical Computing . R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ .
• Tingley, D., Yamamoto, T., Hirose, K., Keele, L., & Imai, K. (2014). Mediation: R package for causal mediation analysis. https://www.jstatsoft.org/article/view/v059i05

Clay Ford Statistical Research Consultant University of Virginia Library March 02, 2018

For questions or clarifications regarding this article, contact  [email protected] .

View the entire collection  of UVA Library StatLab articles, or learn how to cite .

Research Data Services

Want updates in your inbox? Subscribe to our monthly Research Data Services Newsletter!

Related categories:

Addressing Moderated Mediation Hypotheses: Theory, Methods, and Prescriptions

Affiliations.

• 1 a University of Kansas.
• 2 b Northwestern University.
• 3 c The Ohio State University.
• PMID: 26821081
• DOI: 10.1080/00273170701341316

This article provides researchers with a guide to properly construe and conduct analyses of conditional indirect effects, commonly known as moderated mediation effects. We disentangle conflicting definitions of moderated mediation and describe approaches for estimating and testing a variety of hypotheses involving conditional indirect effects. We introduce standard errors for hypothesis testing and construction of confidence intervals in large samples but advocate that researchers use bootstrapping whenever possible. We also describe methods for probing significant conditional indirect effects by employing direct extensions of the simple slopes method and Johnson-Neyman technique for probing significant interactions. Finally, we provide an SPSS macro to facilitate the implementation of the recommended asymptotic and bootstrapping methods. We illustrate the application of these methods with an example drawn from the Michigan Study of Adolescent Life Transitions, showing that the indirect effect of intrinsic student interest on mathematics performance through teacher perceptions of talent is moderated by student math self-concept.

Conducting a moderated mediation analysis

This vignette reports a moderated mediation analysis. This will help getting familiar with the several helpers JSmediation offers to conduct moderated mediation analysis.

Simple mediation refers to the pattern of statistical relationships in which the effect of a variable on another goes through a third variable. Sometimes, this indirect relationship is conditional to a fourth variable—a moderator. In such a case, we talk about moderated mediation . The JSmediation package offers a collection of tools to conduct and report a moderated mediation analysis (Muller et al., 2005) .

Conducting a Moderated Mediation Analysis

The Introduction to JSmediation vignette uses a data set collected by Ho et al. (2017) to illustrate simple mediation. This data set contains the result of an experiment in which Ho et al. focus on hypodescent among Black Americans. To illustrate how JSmediation can be used to conduct a moderated mediation analysis, we will expand on this example.

Hypodescent refers to a rule observed in multiracial categorization. Research shows that majority group social perceivers (i.e., White Americans) typically associate Black-White multiracials with their lower status group rather than with their higher group (i.e., with Black Americans rather than with White Americans). This phenomenon is called the hypodescent rule and it was retrieved in the Black Americans population by Ho et al. (2017) . The hypothesis is that such categorization occurs because of a sentiment of linked fate between Black Americans and Black-White multiracials.

This observation was supported by data showing that Black Americans felt a stronger sentiment of linked fate between them and Black-White multiracials when they read about an article discussing how Black-White multiracials were likely to suffer from discrimination (compared to an article discussing how they were not). The higher sentiment of linked fate was associated with a higher use of the hypodescent rule in multiracial categorization. In other words, the effect of a discrimination condition on hypodescent rule was mediated by the feeling of linked fate.

In the current vignette, we will investigate whether this indirect effect is moderated by social dominance orientation (SDO). Indeed, SDO correlates with preferences for intergroup equality—higher levels indicates resistance to intergroup equality and support for maintaining status quo, lower levels indicates support for intergroup equality and reducing hierarchy (Ho et al., 2017) . Ho et al. (2017) assumed that the indirect effect will be stronger for Black Americans showing lower levels of SDO. They should feel a higher sentiment of linked fate when presented information related to the high discrimination of Black-White multiracials.

While this vignette was written to illustrate features from the JSmediation package, we will use functions from other packages (e.g., dplyr ), but make it explicit when we do so. In other words, rather than calling library(dplyr) to attache the dplyr package and then calling its function (e.g., mutate ), we will use dplyr::mutate , mutate will suffice.

Data Preparation

The data set collected by Ho et al. (2017) contains the information necessary for the test of our hypothesis of moderated mediation. It contains a condition column indicating whether participants were read about the high or low discrimination of Black-White multiracials and a hypodescent column. The data set also contains a linkedfate column indicating the feeling of linked fate participants felt in the experimental situation. The simple mediation analysis in the Introduction to JSmediation vignette shows that participants in high discrimination conditions tend to feel a higher level of linked fate which results in higher hypodescent. We will now investigate whether this indirect effect is more likely to be observed among participants who show higher levels of SDO (thanks to the sdo column).

Before we run the moderated mediation analysis, we must tweak the data set a little bit. Muller et al. (2005) recommend centering the continuous predictors and contrast-coding the categorical ones. JSmediation offers helpers like build_contrast or standardize_variable to do so.

The first function we used, standardize_variable , allows us to center and reduce (to standardize) a set of variables (here: sdo and linkedfate ). The suffix argument of this function is used to name the standardized variables. Here, because we set suffix = "c" , the new variables are named sdo_c and linkedfate_c . When the suffix argument is blank, original variables are overwritten. Because JSmediation was designed to fit within the tidyverse, standardize_variable is especially useful in a dplyr chain .

Here, we also used a second function from JSmediation : build_contrast . This function takes a character vector as a first argument and, then, returns a contrast-coded variables. It is designed for variables with two conditions, which is the case of condition . Because we want to include the transformed variable in ho_et_al , we can use the dplyr::mutate function which creates (or modify) variables . With this method, we contrast-coded the condition variable.

Fitting the moderated mediation model

We will now run our moderated mediation analysis. Remember, we are testing whether participants in high discrimination condition show higher level of hypodescent because of a higher perception of linked fate, and whether this indirect effect is higher among people with high level of SDO. In other words, we are testing whether SDO moderates the mediation of the effect on condition on hypodescent through linked fate perception. To test this moderated mediation, we will use the mdt_moderated function.

Moderated mediation formalization

Muller et al. (2005) provide an accessible introduction on how to conduct a moderated mediation analysis using the joint-significance methods. In a nutshell, one will fit several linear models to the data set, and, depending on the significance of some coefficients, one will be able to assess moderated mediation.

To better understand the process of testing a moderated mediation, we can glance at the equation describing the moderation of the indirect effect of X on Y. Such moderation can be described as follows:

\[ c \times Mod = c' \times Mod + (a \times Mod) \times b + a \times (b \times Mod) \]

with \(c \times Mod\) the total moderation of the indirect effect, \(c' \times Mod\) the moderation of the direct effect, \((a \times Mod) \times b\) , the moderation of the indirect effect passing by the moderation of \(a\) , and \(a \times (b \times Mod)\) , the moderation of the indirect effect passing by the moderation of \(b\) (see Models section; Muller et al., 2005).

Either both \(a \times Mod\) and \(b\) or both \(a\) and \(b \times Mod\) need to be simultaneously significant for the moderation of the indirect effect to be claimed (Muller et al., 2005).

Running the Analysis with mdt_moderated

mdt_moderated works like every mdt_* function. It will be fit the necessary linear models (see the Models section in the help page of mdt_moderated ) and returns the relevant coefficient in a nicely formatted output.

To use the function, we will have to describe the role of the variable in the ho_et_al data set.

Most of the math occurs when mdt_moderated is evaluated. The moderated_mediation_fit that we just computed contains every linear model useful to assess moderated mediation, as described in Muller et al. (2005) .

As we can see, it is possible to see \(a\) , \(a \times Mod\) , \(b\) , and \(b \times Mod\) in the mdt_moderated output.

If a moderated mediation model is significant, Yzerbyt et al. (2018) recommend reporting the moderated mediation index. To compute this index, JSmediation offers the add_index method. In the case of moderated mediation, we will have to specify whether we are in a scenario where we assume that the moderator impacts the relation between or IV and the mediator (i.e., both \(a \times Mod\) and \(b\) are significant), the relation between the moderator and the DV (i.e., both \(a\) and \(b \times Mod\) are significant), or both. To do so, we must use the stage argument.

The output of the new object now contains the index.

Reporting Moderated Mediation

The moderated_mediation_fit_w_index object that we computed contains every bit of information to report a moderated mediation in an analysis, following Yzerbyt et al.’s (2018) recommendations:

To assess whether SDO moderates the indirect effect of discrimination condition on hypodescent through linked fate perception, we conducted a moderated mediation analysis (Muller et al., 2005) . This analysis first revealed that SDO moderates the effect of condition on linked fate perception, t (820) = 4.25, p < .001. It also revealed that the effect of linked fate perception for a mean level of SDO, and controlling for the effect of the condition and for the moderation of the condition by SDO, predicted significantly hypodescent, t (818) = 3.75, p < .001. This pattern reveals the existence of a moderated mediation. We also computed the first stage moderated mediation index consistent with our joint-significant analysis and it confirmed the moderated mediation, -0.0432, CI 95% [-0.0778; -0.0171] (Monte Carlo simulation, 5000 simulations; Yzerbyt et al., 2018).

Compute the indirect effect for a moderator’s value

When reporting a moderated mediation analysis, it is sometimes useful to know what the indirect effect is for several values of the moderator. The compute_indirect_effect_for function serves that purpose.

compute_indirect_effect_for takes as the first argument a moderated mediation model for which we want to compute the indirect effect at a specific value of the moderator and as the second argument the value of the moderator.

In our example, we standardized our moderator. A mean value of the moderator therefore is sdo_c = 0 . If we wanted to compute the effect of condition on hypodescent through linked fate for a mean value of SDO, we would have to run the following:

We could also compute this indirect effect for participants who score one standard deviation above the mean in terms of SDO.

As we can see, the indirect effect estimate decreases when the SDO increases. This result is consistent with the negative value of the moderated mediation index.

Want to create or adapt books like this? Learn more about how Pressbooks supports open publishing practices.

Section 7.3: Moderation Models, Assumptions, Interpretation, and Write Up

Learning Objectives

At the end of this section you should be able to answer the following questions:

• What are some basic assumptions behind moderation?
• What are the key components of a write up of moderation analysis?

Moderation Models 

Difference between mediation & moderation.

The main difference between a simple interaction, like in ANOVA models or in moderation models, is that mediation implies that there is a causal sequence. In this case, we know that stress causes ill effects on health, so that would be the causal factor.

Some predictor variables interact in a sequence, rather than impacting the outcome variable singly or as a group (like regression).

Moderation and mediation is a form of regression that allows researchers to analyse how a third variable effects the relationship of the predictor and outcome variable.

Moderation analyses imply an interaction on the different levels of M

PowerPoint: Basic Moderation Model

Consider the below model:

• Chapter Seven – Basic Moderation Model

Would the muscle percentage be the same for young, middle-aged, and older participants after training? We know that it is harder to build muscle as we age, so would training have a lower effect on muscle growth in older people?

Example Research Question:

Does cyberbullying moderate the relationship between perceived stress and mental distress?

Moderation Assumptions

• The dependent and independent variables should be measured on a continuous scale.
• There should be a moderator variable that is a nominal variable with at least two groups.
• The variables of interest (the dependent variable and the independent and moderator variables) should have a linear relationship, which you can check with a scatterplot.
• The data must not show multicollinearity (see Multiple Regression).
• There should be no significant outliers, and the distribution of the variables should be approximately normal.

Moderation Interpretation

PowerPoint: Moderation menu, results and output

Please have a look at the following link for the Moderation Menu and Output:

• Chapter Seven – Moderation Output

Interpretation

The effects of cyberbullying can be seen in blue, with the perceived stress in green. These are the main effects of the X and M variable on the outcome variable (Y). The interaction effect can be seen in purple. This will tell us if perceived stress is effecting mental distress equally for average, lower than average or higher than average levels of cyberbullying. If this is significant, then there is a difference in that effect. As can be seen in yellow and grey, cyberbullying has an effect on mental distress, but the effect is stronger for those who report higher levels of cyberbullying (see graph).

Moderation Write Up

The following text represents a moderation write up:

A moderation test was run, with perceived stress as the predictor, mental distress as the dependant, and cyberbullying as a moderator.  There was a significant main effect found between perceived stress and mental distress, b = -1.23, BCa CI [1.11, 1.34], z =21.38 , p <.001, and nonsignificant main effect of cyberbullying on mental distress b = 1.05, BCa CI [0.72, 1.38], z=6.28, p < .001. There was a significant interaction found by cyberbullying on perceived stress and mental distress, b = -0.05, BCa CI [0.01, 0.09], z=2.16, p =.031. It was found that participants who reported higher than average levels of cyberbullying experienced a greater effect of perceived stress on mental distress ( b = 1.35, BCa CI [1.19, 1.50], z=17.1, p < .001), when compared to average or lower than average levels of cyberbullying ( b = 1.23, BCa CI [1.11, 1.34], z=21.3, p < .001, b = 1.11, BCa CI [0.95, 1.27], z=13.8, p < .001, respectively). From these results, it can be concluded that the effect of perceived stress on mental distress is partially moderated by cyberbullying.

Statistics for Research Students Copyright © 2022 by University of Southern Queensland is licensed under a Creative Commons Attribution 4.0 International License , except where otherwise noted.

Share This Book

• Course Notes
• Moments, Z-scores, Probability, & Sampling Error
• Hypothesis Testing - z-tests & t-tests
• Power and Effect Size
• Introduction of Analysis of Variance (ANOVA)
• Follow Up to One-way ANOVA
• Multiple Comparisons (Bonferroni vs FDR)
• Calculating the Two-Way Analysis of Variance
• Following up the Two-Way ANOVA
• Paired t-test to RM ANOVA
• RM ANOVA - Two-way, Graphing & Follow ups
• Mixed ANOVA - Two-way, Graphing & Follow ups
• Pearson's Chi-Square and Other Useful Non-Parametrics
• Correlations and Linear Regression
• Partial and Semipartial (part) Correlation
• Multiple Regression
• Stepwise and Hierarchical
• Non-Linear Models
• Interactions and Simple Slopes
• Categorical Variables
• Missing Data
• Generalized Linear Model
• Moderated-Mediation
• Linear Mixed Models
• Exploratory Factor Analysis
• Regression Basics
• MLM Two Levels
• MLM Three Levels
• Repeated Factors
• Maximal Fitting
• Parsimonious Fitting
• Flexible Time
• Growth Curves
• Download Rmd

Moderated Mediation

Quick Review

Moderation (process model 1).

“The moderator function of third variables, which partitions a focal independent variable into subgroups that establish its domains of maximal effectiveness regarding a given dependent variable” - Baron & Kenny, 1986.

For example, is success on a kindergarten entrance exam predicted by time to eat the marshmallow, but moderated by parenting style ?

• Measurement scale from -3 to 3 (permissive to authoritative)

Regression models

Test simple slopes.

• Plot the results of each moderator to help us visualize the results
• We must set our moderator levels (-1SD [Permissive] and +1SD [Authoritative])
• Permissive: Parents.C = -1.7 score
• Authoritative: Parents.C = 1.7 score

So what do these results tell us?

Review Mediation (Process Model 4)

“The mediator function of a third variable, which represents the generative mechanism through which the focal independent variable can influence the dependent variable of interest” - Baron & Kenny, 1986.

For example, do does children respect/trust in of authority figures to keep their promises intermediating in the causal chain for success.

Test mediation

Steps, 1) Y~X [c path], 2) Med~X [a path], 3) Y~X+Med [b & c’ path]

Moderated-Mediated

“…moderated mediation occurs when the strength of an indirect effect depends on the level of some variable, or in other words, when mediation relations are contingent on the level of a moderator” - see Preacher, Rucker, & Hayes (2007)

• Moderation examines how subgroups influence the strength of the relationship between X to Y
• Moderated-Mediated is that effect of the mediator is moderated

You collect 300 4-year-olds, give them the Marshmallow test (measure their time to eat the marshmallow). You operationalize success as how they did at the end of the year on kindergarten entrance exam. - You also measure how must trust they have in authority figures (proposed Mediator). - You also measure the parenting style of the high warmth parents (proposed Moderator) - Permissive parents (no boundaries) vs Authoritative (strict) - Measurement scale from -3 to 3 (permissive to authoritative) - What if trust is moderated by parenting styles - We expect mediator ( trust ) to be stronger for kids with authoritative than permissive parents ( indirect path ) - But it could also be possible that children from authoritative parents who last longer will show more success ( direct path )

Model approaches

There are multiple ways to think what about testing a moderated mediation. Preacher, Rucker, & Hayes (2007) argue you can test the moderation on a, b, c’ pathways directly and that should be theorized moderation based on where you theorize it to be. Imai et al, (2010) have proposed a different framework that allows for generalized approach, but they instead of think about it moderating the direct and indirect path (not a or b)

Only the indirect pathway is moderated

Both a and b path are moderated

Process Model 58

Just a path is moderated

Process Model 7

Just b path is moderated

Process Model 14

Direct and indirect pathway are moderated

Simulation below:

Moderation on Path a, b, c’

• In this case, we think the direct (c’) and indirect path (a,b) is moderated by Parenting style. ( Note this is Hayes’ process model 59 )

So we have to interact it in Model 2 (Mediator Model) and Model 3 (Outcome Model)

Model 2 from Kenny becomes, \(M ~ X*Mod\)

Model 3 from Kenny becomes, \(Y ~ M*Mod+X*Mod\)

Since we are going to be dealing with interactions, we should center our scores before analysis

• Regression models:
• Mediator and Outcome Models ( Note difference in DV when reading table ):

Test Moderation on Path a, b, c’

• Again, we can use the +/- 1SD from the mean (also we can use zero if wanted to see the average parent)
• This allows us to view the impact of the moderator on the direct and indirect effect

Permissive Parents

• We set the covariates = list(Parents.C = Permissive) in our mediation (remember we defined this above as having a Parents.C = -1.7 score)

• For children of permissive parents, they are more successful if they have more trust

Authoritative Parents

• We set the covariates = list(Parents.C = Authoritative) in our mediation (remember we defined this above as having a Parents.C = 1.7 score)

• For children of authoritative parents, they are more successful if they have more trust

Difference Direct/indirect effect

• we must first run a third moderation (not accoutining for the covariate)
• use the test.modmed function at testing between the covariates, covariates.1 = list(Parents.C = Permissive) and covariates.2 = list(Parents.C = Authoritative)
• Given the we did not have a direct effect to begin with, the difference in direct effect levels is not interesting
• The negative value is that indirect effect for Permissive ACME) - Authoritative ACME = -7.12 and it was significantly different from 0. So the effect is bigger for Authoritative parenting as we predicted.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of personality and social psychology , 51(6), 1173.

Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007). Addressing moderated mediation hypotheses: Theory, methods, and prescriptions. Multivariate behavioral research, 42(1), 185-227.

Imai, K., Keele, L., & Tingley, D. (2010). A general approach to causal mediation analysis. Psychological methods, 15(4), 309.

• Help & FAQ

Addressing moderated mediation hypotheses: Theory, methods, and prescriptions

Research output : Contribution to journal › Article › peer-review

This article provides researchers with a guide to properly construe and conduct analyses of conditional indirect effects, commonly known as moderated mediation effects. We disentangle conflicting definitions of moderated mediation and describe approaches for estimating and testing a variety of hypotheses involving conditional indirect effects. We introduce standard errors for hypothesis testing and construction of confidence intervals in large samples but advocate that researchers use bootstrapping whenever possible. We also describe methods for probing significant conditional indirect effects by employing direct extensions of the simple slopes method and Johnson-Neyman technique for probing significant interactions. Finally, we provide an SPSS macro to facilitate the implementation of the recommended asymptotic and bootstrapping methods. We illustrate the application of these methods with an example drawn from the Michigan Study of Adolescent Life Transitions, showing that the indirect effect of intrinsic student interest on mathematics performance through teacher perceptions of talent is moderated by student math self-concept.

ASJC Scopus subject areas

• Experimental and Cognitive Psychology
• Arts and Humanities (miscellaneous)
• Statistics and Probability

Access to Document

• 10.1080/00273170701341316

Other files and links

• Link to publication in Scopus

Fingerprint

• hypothesis INIS 100%
• mediation INIS 100%
• Indirect Effect Mathematics 100%
• Mathematics Performance Psychology 100%
• Teachers Psychology 100%
• Adolescents Psychology 100%
• Self-Concept Psychology 100%
• Conditionals Mathematics 75%

T1 - Addressing moderated mediation hypotheses

T2 - Theory, methods, and prescriptions

AU - Preacher, Kristopher J.

AU - Rucker, Derek D.

AU - Hayes, Andrew F.

N1 - Funding Information: This work was funded in part by National Institute on Drug Abuse Grant DA16883 awarded to the first author while at the University of North Carolina at Chapel Hill. We thank Li Cai for valuable input regarding derivations in the Technical Appendix, Stephanie Madon and Courtney Stevens for help with the applied example, and Daniel J. Bauer for helpful advice in improving the manuscript. The SPSS macro syntax is available online through http://www.quantpsy.org/.

N2 - This article provides researchers with a guide to properly construe and conduct analyses of conditional indirect effects, commonly known as moderated mediation effects. We disentangle conflicting definitions of moderated mediation and describe approaches for estimating and testing a variety of hypotheses involving conditional indirect effects. We introduce standard errors for hypothesis testing and construction of confidence intervals in large samples but advocate that researchers use bootstrapping whenever possible. We also describe methods for probing significant conditional indirect effects by employing direct extensions of the simple slopes method and Johnson-Neyman technique for probing significant interactions. Finally, we provide an SPSS macro to facilitate the implementation of the recommended asymptotic and bootstrapping methods. We illustrate the application of these methods with an example drawn from the Michigan Study of Adolescent Life Transitions, showing that the indirect effect of intrinsic student interest on mathematics performance through teacher perceptions of talent is moderated by student math self-concept.

AB - This article provides researchers with a guide to properly construe and conduct analyses of conditional indirect effects, commonly known as moderated mediation effects. We disentangle conflicting definitions of moderated mediation and describe approaches for estimating and testing a variety of hypotheses involving conditional indirect effects. We introduce standard errors for hypothesis testing and construction of confidence intervals in large samples but advocate that researchers use bootstrapping whenever possible. We also describe methods for probing significant conditional indirect effects by employing direct extensions of the simple slopes method and Johnson-Neyman technique for probing significant interactions. Finally, we provide an SPSS macro to facilitate the implementation of the recommended asymptotic and bootstrapping methods. We illustrate the application of these methods with an example drawn from the Michigan Study of Adolescent Life Transitions, showing that the indirect effect of intrinsic student interest on mathematics performance through teacher perceptions of talent is moderated by student math self-concept.

UR - http://www.scopus.com/inward/record.url?scp=34547179995&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34547179995&partnerID=8YFLogxK

U2 - 10.1080/00273170701341316

DO - 10.1080/00273170701341316

M3 - Article

C2 - 26821081

AN - SCOPUS:34547179995

SN - 0027-3171

JO - Multivariate Behavioral Research

JF - Multivariate Behavioral Research

BRIEF RESEARCH REPORT article

Self-esteem and problematic smartphone use among adolescents: a moderated mediation model of depression and interpersonal trust.

• 1 Department of Psychology, Renmin University of China, Beijing, China
• 2 School of Journalism and Communication, Renmin University of China, Beijing, China

Research has found that self-esteem is negatively associated with problematic smartphone use (PSU). However, the internal mechanisms underlying that relationship need further investigation. The purpose of this study was to investigate the roles of depression and interpersonal trust in the relationship between self-esteem and PSU among adolescents. A questionnaire comprised of the Rosenberg Self-esteem Scale, Inclusive General Trust Scale (IGTS), Self-rating Depression Scale (SDS), and personal questions was administered to 637 students (female = 355) at two middle schools in Shanghai, China. Correlation analyses, mediation analysis, and moderated mediation analysis were performed. A moderated mediation model was established, which revealed: (1) a significant negative association between self-esteem and PSU, (2) depression mediated the relationship between self-esteem and PSU, and (3) the influence of depression on the relationship between self-esteem and PSU was moderated by interpersonal trust. The results indicated that low self-esteem was a risk factor, and interpersonal trust was a moderating factor for PSU among adolescents in the sample. Building adolescents’ self-esteem and increasing their interpersonal trust might decrease their PSU.

Introduction

The past decade has witnessed a rapid rise in smartphone use. As of December 2018, the number of Chinese youth netizens exceeded 200 million, of which 90% were smartphone users ( China Internet Network Information Center, 2019 ). Particularly among adolescents, the smartphone, as an internet terminal, is the main point of access to the internet ( We Are Social and Hootsuite, 2018 ; China Internet Network Information Center, 2019 ). However, despite their portability, convenience, and versatility, smartphones are causing some problems, one of which is referred to as “problematic smartphone use” (PSU) ( Kuss et al., 2018 ; Xie et al., 2018 ). PSU is defined as inappropriate or excessive uses of smartphones that might interfere with everyday life, impair social functions, and/or lead to psychological and/or behavioral problems ( Billieux et al., 2015 ). The concept of PSU mainly derives from two similar concepts. “Problematic mobile phone use” is a term and concept found in recent literature ( De-Sola et al., 2017 ; Jiang and Zhao, 2017 ) and reports published before the smartphone boom ( Bianchi and Phillips, 2005 ; Jenaro et al., 2007 ; Chóliz, 2010 ), and “problematic internet use” or “internet addiction” is a concept that has provided theoretical and methodological support for studies on PSU ( Salehan and Negahban, 2013 ; Lee et al., 2014 ; Lin et al., 2014 ).

Many studies have found adverse consequences of PSU, particularly for adolescents. It has a negative influence on their physical health, such as headaches ( Söderqvist et al., 2008 ), sleep problems ( Söderqvist et al., 2008 ; Lemola et al., 2015 ; Xie et al., 2018 ), eye and vision problems ( Xie et al., 2018 ), and so on. Other studies found that PSU damaged adolescents’ psychological and social functions ( Yen et al., 2009 ), and it led to behavioral problems, such as aggression and smoking ( Augner and Hacker, 2012 ). Furthermore, PSU might be associated with concentration difficulties ( Söderqvist et al., 2008 ) and poor academic performance ( Hawi and Samaha, 2016 ; Samaha and Hawi, 2016 ; Yang et al., 2019 ).

Because PSU might impede adolescents’ healthy development, concerns are growing about the possible correlates of PSU, such as various features of personality, emotion, interpersonal relationships, and so on ( Bianchi and Phillips, 2005 ; Butt and Phillips, 2008 ; De-Sola et al., 2017 ; Wang et al., 2017 ; Kim and Koh, 2018 ). Some scholars proposed that PSU is an addiction-like phenomenon, similar to problem gambling ( Billieux et al., 2015 ; Wang et al., 2016 ; Elhai and Contractor, 2018 ). Therefore, many studies have adopted the behavioral addiction perspective ( Jiang and Zhao, 2017 ; Ihm, 2018 ). Because adolescents are in a critical period of physical and psychological development, exploring the causes of PSU helps us better mitigate smartphones’ adverse consequences.

Self-esteem deserves attention as a possible cause. Self-esteem is about mental representations of the self-regarding overall feelings of self-worth and self-acceptance ( Rosenberg, 1965 ), and self-esteem is essential to adolescents’ social development and social adaptation ( Robins and Trzesniewski, 2005 ; Trzesniewski et al., 2006 ). Based on the ecological systems theory and dual-factor resiliency theory, an explanatory model of adolescent problem behavior recognizes low self-esteem as one of the individual-level risk factors ( Jessor et al., 2003 ). Many previous studies have found robust associations between low self-esteem and behavioral problems or deviance, and it has been related to aggression ( Donnellan et al., 2005 ; Garofalo et al., 2016 ), smoking ( Carters and Byrne, 2013 ), alcohol and/or drug use ( Wild et al., 2004 ; Kavas, 2009 ), problem gambling ( Rodda et al., 2004 ), and delinquency ( Donnellan et al., 2005 ). Meanwhile, a cognitive-behavioral model of internet addiction also suggests that low self-esteem is a critical aspect of maladaptive cognition of the self, which is the core factor leading to problematic internet use ( Davis, 2001 ). Of particular relevance to the current study, research has linked self-esteem and PSU such that low self-esteem is a risk factor of PSU ( Butt and Phillips, 2008 ; Ha et al., 2008 ; Billieux, 2012 ; Isiklar et al., 2013 ), and self-esteem often has been identified as an antecedent of PSU. Wang et al. (2017) suggested that low-quality peer relationships among adolescents might lead to low self-esteem, which, in turn, might lead to PSU. Kim and Koh (2018) also found that low self-esteem was associated with PSU.

Although many studies have found direct relationships between self-esteem and PSU, more research about the internal mechanisms is needed. First, a systematic review pointed out that the link between self-esteem and PSU found by previous studies is inconsistent, such that the bivariate effects were small or moderately strong and the effects found in multivariate analyses were weaker or non-significant ( Elhai et al., 2017 ). Therefore, exploring mediating and moderating factors might shed light on the relationship. In addition, a better understanding of mediating and moderating influences might improve our understanding of the etiology of PSU and support development of effective interventions.

Davis’s (2001) cognitive-behavioral model proposes that psychological distress, such as depression, is an essential and significant catalyst of problematic internet use. Recent studies on PSU indicated that depression predicted PSU ( Panova and Lleras, 2016 ; Kuss et al., 2018 ), and it has been linked to problematic internet use, problem gambling, and other behavioral problems ( Armstrong et al., 2000 ; Rodda et al., 2004 ). Kim et al. (2015) suggested that, when individuals felt depressed, they tend to use their smartphones to cope with their negative emotions. In other words, using a smartphone is considered an experiential avoidance strategy to divert aversive emotional content. However, experiential avoidance is ineffective for achieving that outcome, and, instead, it has adverse consequences ( Kim et al., 2015 ; Chou et al., 2018 ). Meanwhile, influential theories of depression have considered aspects of low self-esteem as a vulnerability factor that confers risk to depression. For instance, Orth and Robins (2013) validated the vulnerability model of the relationship between low self-esteem and depression, which considered self-esteem to be a leading cause of depression. Further, individuals with lower self-esteem was more likely to develop depression than those with high self-esteem ( Orth et al., 2014 ). From the view of the buffer hypothesis, low self-esteem is also one of the most important susceptibility factors to depression, which tends to lead to depression in individuals under the influence of stressful life events ( Murrell et al., 1991 ; Hankin et al., 2007 ). In particular, the hopelessness and self-esteem theory of depression holds that depression occurs as a result of negative attributional style, low self-esteem and negative life events, where hopelessness plays a mediating role, low self-esteem playing a stimulating role, and high self-esteem acting as a buffer ( Metalsky et al., 1993 ). A meta-analysis of longitudinal studies supported this contention, revealing that the effect of self-esteem on depression was significantly stronger than that of depression on self-esteem ( Sowislo and Orth, 2013 ). This implies that depression might play a mediating role in the relationship between self-esteem and PSU.

Interpersonal trust is vital to healthy psychosocial development and key to the formation and maintenance of healthy interpersonal relationships ( Gurtman, 1992 ; Simpson, 2007 ). Interpersonal trust has been defined as a psychological state of voluntarily placing oneself in an undefended or vulnerable position based on confident expectations of the good intentions and actions of others ( Rousseau et al., 1998 ; Haselhuhn et al., 2015 ). The emancipation theory of trust proposes that people with high levels of interpersonal trust often expand their social networks ( Yamagishi and Yamagishi, 1994 ; Yamagishi, 1998 ). Previous studies also found that interpersonal trust among children and adolescents related to higher social status ( Buzzelli, 1988 ), less loneliness ( Rotenberg et al., 2004 ), and better-quality peer relationships ( Rotenberg et al., 2004 ). Ihm (2018) interpreted PSU as a social problem stemming from a lack of offline social networks and in-person social support. When an individual’s needs are not being met, she or he might turn to virtual sources of support as an alternative, which might lead to pathological smartphone usage. Davis’s (2001) cognitive-behavioral model also proposes that the need for social contact and reinforcement obtained online would increase the desire to remain in a virtual social life. From this perspective, the benefits of interpersonal trust mentioned above might reduce the emergence of PSU by providing alternative ways to meet individuals’ needs.

Previous studies have also found a negative association between trust and depression ( Kim et al., 2012 ; Betts et al., 2017 ). Specifically, lower interpersonal trust increased the incidence of depression ( Kim et al., 2012 ; Frank et al., 2014 ; Betts et al., 2017 ), whereas higher interpersonal trust (as a component of social capital) buffered the effects of financial stress on depression ( Frank et al., 2014 ). Moreover, individuals with higher interpersonal trust had better psychosocial adjustment ( Lester and Gatto, 1990 ; Rotenberg et al., 2004 ). A study of Chinese and American adolescents and young adults found that interpersonal trust facilitated the support-seeking process and was associated with appropriate help-seeking, which, in turn, predicted the likelihood of expressing emotional distress to friends ( Mortenson, 2009 ). In this way, trust might also mitigate the adverse effects of depression. Therefore, interpersonal trust might buffer the relationship between self-esteem and PSU via depression.

This study’s main objectives were to explore the relationships among self-esteem, depression, interpersonal trust, and PSU and to test a moderated mediation model of the influence of self-esteem on PSU mediated by depression and moderated by interpersonal trust. Based on the results of previous studies, three hypotheses were proposed as follows.

Hypothesis 1. Self-esteem would be negatively related to problematic smartphone use.

Hypothesis 2. Depression would mediate the effect of self-esteem on problematic smartphone use.

Hypothesis 3. Interpersonal trust would moderate the direct and indirect relationships between self-esteem and PSU via depression. Specifically, interpersonal trust would buffer the direct effect of self-esteem on problematic smartphone use (Hypothesis 3a), and buffer the mediating influence of depression on the effect of self-esteem on problematic smartphone use (Hypothesis 3b).

Materials and Methods

Participants and procedures.

The data for this study were collected by trained and experienced research assistants in the classrooms of two middle schools in Shanghai, China, in May 2016. The respondents completed an anonymous questionnaire comprising several self-report inventories and personal questions. A token gift notebook was given to each respondent. Initially, 689 adolescents were asked to participate. Of them, 52 did not use a smartphone or did not answer all the questions, and they were dropped from the analysis. The valid sample used in the analysis comprised 637 respondents (92% response rate) ( M age = 15.38 years, SD = 1.29 years), of which 355 (55.7%) were female.

Problematic smartphone use was measured by a 33-item inventory developed by Xie et al. (2018) . This inventory covers four aspects of PSU: overuse, withdrawal, compulsive behavior, and disturbances. The items describe problems, obsession, or dysfunction related to smartphone use, such as “I feel restless and irritable when the smartphone is unavailable” and “My grades or school work suffer because of overuse of the smartphone.” Adolescents were asked to report how often they experienced these problems on a four-point Likert-type scale where 1 = never through 4 = always . The scale demonstrated strong internal reliability in the current study (α = 0.95).

Self-Esteem

The Rosenberg Self-Esteem Scale ( Rosenberg, 1965 ) was used to measure self-esteem. The scale includes 10 items (e.g., “I am able to do things as well as most other people”). The respondents rated each item on a five-point Likert-type scale where 1 = not very true of me through 4 = very true of me , with higher scores representing higher self-esteem. Cronbach’s α of the scale in this study was 0.82.

Depression was assessed by the Self-rating Depression Scale (SDS; Zung, 1965 ), which consists of 20 items [e.g., “I feel down hearted and blue” and “I feel hopeful about the future” (reverse-coded)]. Each item was scored on a four-point scale where 1 = never through 4 = always and higher scores indicated higher depression. In the current study, Cronbach’s α for the SDS was 0.79.

Interpersonal Trust

Interpersonal trust was measured using the Inclusive General Trust Scale (IGTS), a nine-item scale developed by Yamagishi et al. (2015) . This scale captures the belief aspect (e.g., “Generally, I trust others”) and the preference aspect (e.g., “I hate to lose because of having counted on someone” [reverse-coded]) of trust. Cronbach’s α for the IGTS in the current study was 0.80.

First, the research assistants explained the study and the questionnaire to the prospective respondents in the classrooms. Second, the respondents gave their written informed consent. Third, the questionnaire was distributed, and the adolescents completed it. The research assistants remained in the classrooms to answer any questions that they might have had. It took about 15 min for all of the adolescents to complete the questionnaire, which was collected at that time. Last, the respondents were debriefed and received their token of appreciation for participating in the study.

Statistical Analyses

The analyses were performed using IBM SPSS software. The mediation model (Model 4) and the moderated mediation model (Model 59) were tested using the PROCESS macro ( Hayes, 2013 ). Age and gender were included in both models as control variables. The indirect effects were tested with bias-corrected bootstrapping ( n = 5,000) and 95% confidence intervals (CI) for the indices. When a 95% bootstrapped CI does not include zero, it indicates the parameter is statistically significant.

This study was conducted with the approval of the Research Ethics Review Committee of the author’s institution. All participants were well informed in advance and debriefed at the end.

Descriptive Statistics

Means, standard deviations, and Pearson’s correlations ( r ) were calculated on all the study variables ( Table 1 ). As expected, self-esteem and PSU were negatively correlated, depression negatively correlated with self-esteem, and depression positively correlated with PSU. Interpersonal trust positively correlated with self-esteem, and it was negatively correlated with PSU and with depression. All bivariate correlations were statistically significant ( p < 0.001). No independent samples t -test results on the gender differences in the variables were statistically significant.

Table 1. Descriptive statistics and correlations among variables.

Mediating Effect of Depression

Hypothesis 1 and 2 was tested controlling for the effects of age and gender ( Table 2 ). The total effects of self-esteem on PSU were statistically significant (β = −0.22, SE = 0.04, p < 0.001), indicating that the respondents with lower self-esteem had higher PSU, which supported Hypothesis 1. Although self-esteem had no direct effect on PSU, it directly influenced depression, and the association of depression with PSU also was significant. The standardized indirect effect of self-esteem on PSU via depression was significant, indirect effect = −0.18, SE = 0.03, 95% CI = [−0.24, −0.12], and the indirect effect’s proportion of the total effect was 82.7%. In support of Hypothesis 2, depression had a mediating effect on the relationship of self-esteem to PSU through its negative relationship with self-esteem (depression increased as self-esteem decreased), which, in turn, related to higher PSU.

Table 2. Testing the mediating effect of self-esteem on PSU.

Moderated Mediation Effects

Hypotheses 3a and 3b were tested by estimating a moderated mediation model (model 59) with PROCESS macro ( Hayes, 2013 ) that included age and gender as control variables.

Table 3 indicates there was a significant negative direct influence of self-esteem on depression; however, its direct effect on PSU was not statistically significant. A significant moderating effect of interpersonal trust on the direct effect of self-esteem on PSU was found ( Table 3 ). We plotted the results for PSU predicted by self-esteem separately for low (one standard deviation below the mean) and high (one standard deviation above the mean) interpersonal trust ( Supplementary Figure 1 ). Simple slope tests revealed that respondents with high trust and relatively high self-esteem had significantly lower PSU, β simple = −0.16, SE = 0.07, p = 0.02, 95% CI = [−0.30, −0.02]. However, for the respondents with low trust, the direct effect of self-esteem on PSU was not statistically significant, β simple = 0.05, SE = 0.06, p = 0.43, 95% CI = [−0.07, 0.17].

Table 3. Testing the moderated mediating effect of self-esteem on PSU.

The direct effect of depression on PSU was statistically significant ( Table 2 ). However, the relationship between depression and PSU was moderated by interpersonal trust. Simple slope tests indicated that the association of depression with PSU was weaker for respondents with high trust (i.e., one standard deviation above the mean; β simple = 0.16, p = 0.03, 95% CI = [0.01, 0.30]) than for respondents with low trust (i.e., one standard deviation below the mean; β simple = 0.39, SE = 0.06, p < 0.001, 95% CI = [0.26, 0.51]; Supplementary Figure 2 ).

Last, the bias-corrected percentile bootstrap method further revealed a significant moderated mediation effect, β = 0.12, SE = 0.05, 95% CI = [0.02, 0.22], in which interpersonal trust moderated the mediating effect of depression by buffering its influence on PSU. The indirect effect of self-esteem on PSU via depression was statistically significant for the respondents with low trust (i.e., one standard below the mean), β = −0.20, SE = 0.04, 95% CI = [−0.27, −0.13]. In contrast, this indirect effect was non-significant for respondents with high trust (i.e., one standard above the mean), β = −0.08, SE = 0.04, 95% CI = [−0.16, 0.0001]. The moderated mediation model is shown in Figure 1 . Given that interpersonal trust moderated just the second stage of the mediation process, a second-stage moderation model, which is a type of moderated mediation model, was established ( Hayes, 2013 ). Therefore, Hypothesis 3 was partially supported.

Figure 1. The moderated mediation model controlling for age and gender (all coefficients standardized). SE, self-esteem; DEP, depression; PSU, problematic smartphone use; IT, interpersonal trust. * p < 0.05; *** p < 0.001.

Empirical support exists for a link between self-esteem and PSU in which self-esteem negatively predicts PSU. However, its internal mechanisms are ambiguous. The purpose of this study was to investigate whether self-esteem indirectly related to PSU via depression and whether the direct and indirect associations between self-esteem and PSU were moderated by interpersonal trust. The findings indicated that the predictive effect of low self-esteem on PSU was mostly explained by increased depression, and it was buffered by interpersonal trust. For respondents with high trust, higher self-esteem was associated with less PSU; however, for respondents with low trust, the direct effect of self-esteem on PSU was not significant. Further, the association between depression and PSU was weaker for the high trust rather than the low trust respondents. Thus, these analyses established a moderated mediation model.

Relationship Between Self-Esteem and PSU

We found a significant total effect of self-esteem on PSU, which supported the findings of previous studies ( Butt and Phillips, 2008 ; Ha et al., 2008 ; Billieux, 2012 ; Isiklar et al., 2013 ). Davis’s (2001) cognitive-behavioral model proposes that maladaptive cognition of the self is the core factor leading to problematic internet use, and low self-esteem certainly is a critical aspect of maladaptive cognition. Individuals with low self-esteem might believe that they are held in high regard only in online interactions, and, consequently, they use smartphones to obtain approval and recognition from others. Further, according to the terror management theory of self-esteem ( Greenberg et al., 1997 ), when individuals encounter threats to the development of self-esteem, their need for self-protection is compensated through particular channels. In this way, problem behaviors, including PSU, seem to be compensatory. With the help of various smartphone functions, individuals might meet their needs for esteem through channels not available in the physical world, which, in turn, creates dependence on smartphones.

The Mediating Role of Depression

We developed a mediation model of depression to examine the indirect influence between self-esteem and PSU. As hypothesized, depression mediated the influence of self-esteem on PSU; however, the direct association between self-esteem and PSU was not statistically significant. This finding might be because of the relatively strong link we found between self-esteem and depression, which supports previous studies ( Smetaniuk, 2014 ; Elhai et al., 2017 ). Because no other intermediate variables were analyzed, other factors in this indirect path of depression could not be compared. In the current study, depression largely explained the predictive effect of self-esteem on PSU, which supports previous literature ( Elhai et al., 2017 ).

This study again supports the hopelessness and self-esteem theory and the vulnerability model of depression that low self-esteem affects depression ( Metalsky et al., 1993 ; Orth and Robins, 2013 ). From this perspective, life events can undermine the individual’s psychological protection system through the loss of self-esteem, leading to depression. This study also supports the affective component of PSU causes, that depression plays an important role in the occurrence of PSU.

The Moderating Role of Interpersonal Trust

Interpersonal trust is prominent in theoretical models of lifelong development; its importance has become self-evident. However, few previous studies have analyzed the role of interpersonal trust in the development of PSU or related behavioral problems. Our study identified a second-stage moderation model regarding interpersonal trust and PSU, which enriches the literature of relevant fields. The current study’s adolescent respondents with high interpersonal trust reported relatively good psychosocial adjustment, which aligns with previous studies’ results ( Lester and Gatto, 1990 ; Rotenberg et al., 2004 ). Specifically, the buffering effect of interpersonal trust suggests two dimensions. First, according to the emancipation theory of trust ( Yamagishi and Yamagishi, 1994 ; Yamagishi, 1998 ), individuals with relatively high trust have relatively stable social networks and more social support, which mitigate the negative effects of low self-evaluations and negative emotions. Second, individuals with high interpersonal trust have people they trust to confide in when they feel depressed, which, in turn, reduces the likelihood of PSU.

However, interpersonal trust did not moderate the direct relationship between self-esteem and depression. This was probably because low self-esteem was directly related to increased depression. In a near-infrared spectroscopy study of social pain, Yanagisawa et al. (2011) proposed that general trust and self-esteem, as two psychosocial resources, function at different times during a series of adaptive processes, which reminds us that trust and self-esteem might separately influence depression. Further research is needed to understand the underlying reasons for such findings.

Limitations and Implications

Several limitations should be noted when interpreting this study’s findings. First, all of the measures were self-reports, which might have influenced the study’s validity. Future studies could employ other methods, such as measuring self-esteem using the implicit association test (e.g., Greenwald and Farnham, 2000 ), collecting objective behavioral data on PSU (e.g., Elhai and Contractor, 2018 ), and applying other-report measures to reduce potential common method bias. Second, the study design was cross-sectional, eliminating the possibility of causal inferences ( West, 2011 ). Longitudinal designs should be used for future studies to explore change in the relevant factors and establish temporal order. Third, variables on online activities might suggest interpretations from a different perspective. Future studies that examine the influences of variables, such as online interpersonal trust, might further our understanding of the similarities in and differences between online and in-person social relations/interactions (e.g., Young and Tseng, 2008 ).

Limitations aside, this study extends Davis’s (2001) model to the PSU field and contributes to the understanding of the etiology of PSU and behavioral addiction. This study highlights the affective component of PSU causes and takes interpersonal trust into account, enriching the ecological model of PSU. Furthermore, the findings of this study could help to guide targeted interventions for PSU in adolescents. First, in the family and school education practice, given the characteristics of adolescent self-esteem development, attention should be paid to enhancing and maintaining self-esteem, such as by attribution training ( Metalsky et al., 1993 ), group counseling ( Mackeen and Herman, 1974 ), and family based intervention ( Danielsen et al., 2013 ), thereby reducing the susceptibility to depression, and thus reducing PSU. Second, given the role of interpersonal trust, establishing sound social networks and positive interpersonal interaction, especially peer interaction, thereby increasing interpersonal trust, would also reduce the emergence of PSU. Social cognitive training ( Barrett et al., 1999 ) would help in this way. Third, on such an issue of PSU, not only the cognitive issues should be taken into consideration, but also the affective issues. Effective social support and psychological guidance would also help adolescents to use new technologies rationally and properly.

This study tested a moderated mediation model to examine the psychological factors underlying the relationship between self-esteem and PSU. In brief, the results found that low self-esteem was a risk factor for problematic use of smartphones in a sample of Chinese middle school students, and it predicted PSU through the level of depression, the effects of which on PSU were buffered by interpersonal trust. These findings substantially contribute to our understanding of PSU and behavioral addiction.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

Ethics Statement

The studies involving human participants were reviewed and approved by the Research Ethics Committee of Renmin University of China. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

Author Contributions

CL and YD analyzed the data and wrote the draft. DL and YD collected the data, supervised the study, revised the manuscript, and provided the funding sources. All authors designed the study, had full access to all the data, and have taken responsibility for the integrity of the data and accuracy of the data analysis.

This work was supported by the National Natural Science Foundation of China (31500905), the National Social Science Fund of China (19BSH130), the Fundamental Research Funds for the Central Universities and the Research Funds of Renmin University of China (19XNLG20), and the project of Journalism and Marxism Research Center, Renmin University of China (RMXW2018A005).

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

The authors would like to thank the editor and reviewers for their insightful suggestions for the revision of the manuscript. The authors would also like to thank Editage ( www.editage.cn ) for English language editing.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpsyg.2019.02872/full#supplementary-material

Armstrong, L., Phillips, J. G., and Saling, L. L. (2000). Potential determinants of heavier internet usage. Int. J. Hum. Comput. Stud. 53, 537–550. doi: 10.1006/ijhc.2000.0400

CrossRef Full Text | Google Scholar

Augner, C., and Hacker, G. W. (2012). Associations between problematic mobile phone use and psychological parameters in young adults. Int. J. Public Health 57, 437–441. doi: 10.1007/s00038-011-0234-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Barrett, P. M., Webster, H. M., and Wallis, J. R. (1999). Adolescent self-esteem and cognitive skills training: a school-based intervention. J. Child Fam. Stud. 8, 217–227. doi: 10.1023/a:1022044119273

Betts, L. R., Houston, J. E., Steer, O. L., and Gardner, S. E. (2017). Adolescents’ experiences of victimization: the role of attribution style and generalized trust. J. Sch. Viol. 16, 25–48. doi: 10.1080/15388220.2015.1100117

Bianchi, A., and Phillips, J. G. (2005). Psychological predictors of problem mobile phone use. CyberPsychol. Behav. 8, 39–51. doi: 10.1089/cpb.2005.8.39

Billieux, J. (2012). Problematic use of the mobile phone: a literature review and a pathways model. Curr. Psychiatry Rev. 8, 299–307. doi: 10.2174/157340012803520522

Billieux, J., Maurage, P., Lopez-Fernandez, O., Kuss, D. J., and Griffiths, M. D. (2015). Can disordered mobile phone use be considered a behavioral addiction? An update on current evidence and a comprehensive model for future research. Curr. Addict. Rep. 2, 156–162. doi: 10.1007/s40429-015-0054-y

Butt, S., and Phillips, J. G. (2008). Personality and self reported mobile phone use. Comput. Hum. Behav. 24, 346–360. doi: 10.1016/j.chb.2007.01.019

Buzzelli, C. A. (1988). The development of trust in children’s relations with peers. Child Study J. 18, 33–46. doi: 10.1111/j.1365-2214.1988.tb00563.x

Carters, M. A., and Byrne, D. G. (2013). The role of stress and area-specific self-esteem in adolescent smoking. Aust. J. Psychol. 65, 180–187. doi: 10.1111/ajpy.12019

China Internet Network Information Center, (2019). Statistical Report on Internet Development in China. Available at: http://www.cac.gov.cn/wxb_pdf/0228043.pdf (accessed August 2, 2019).

Google Scholar

Chóliz, M. (2010). Mobile phone addiction: a point of issue. Addiction 105, 373–374. doi: 10.1111/j.1360-0443.2009.02854.x

Chou, W., Yen, C., and Liu, T. (2018). Predicting effects of psychological inflexibility/experiential avoidance and stress coping strategies for internet addiction, significant depression, and suicidality in college students: a prospective study. Int. J. Environ. Res. Public Health 15:788. doi: 10.3390/ijerph15040788

Danielsen, Y. S., Nordhus, I. H., Júlíusson, P. B., Mæhle, M., and Pallesen, S. (2013). Effect of a family-based cognitive behavioural intervention on body mass index, self-esteem and symptoms of depression in children with obesity (aged 7–13): a randomised waiting list controlled trial. Obesity Res. Clin. Pract. 7, e116–e128. doi: 10.1016/j.orcp.2012.06.003

Davis, R. A. (2001). A cognitive-behavioral model of pathological internet use. Comput. Hum. Behav. 17, 187–195. doi: 10.1016/s0747-5632(00)00041-8

De-Sola, J., Talledo, H., Rubio, G., and de Fonseca, F. R. (2017). Psychological factors and alcohol use in problematic mobile phone use in the Spanish population. Front. Psychiatry 8:11. doi: 10.3389/fpsyt.2017.00011

Donnellan, M. B., Trzesniewski, K. H., Robins, R. W., Moffitt, T. E., and Caspi, A. (2005). Low self-esteem is related to aggression, antisocial behavior, and delinquency. Psychol. Sci. 16, 328–335. doi: 10.2307/40064223

Elhai, J. D., and Contractor, A. A. (2018). Examining latent classes of smartphone users: relations with psychopathology and problematic smartphone use. Comput. Hum. Behav. 82, 159–166. doi: 10.1016/j.chb.2018.01.010

Elhai, J. D., Dvorak, R. D., Levine, J. C., and Hall, B. J. (2017). Problematic smartphone use: a conceptual overview and systematic review of relations with anxiety and depression psychopathology. J. Affect. Disord. 207, 251–259. doi: 10.1016/j.jad.2016.08.030

Frank, C., Davis, C. G., and Elgar, F. J. (2014). Financial strain, social capital, and perceived health during economic recession: a longitudinal survey in rural Canada. Anxiety Stress Coping 27, 422–438. doi: 10.1080/10615806.2013.864389

Garofalo, C., Holden, C. J., Zeigler-Hill, V., and Velotti, P. (2016). Understanding the connection between self-esteem and aggression: the mediating role of emotion dysregulation. Aggress. Behav. 42, 3–15. doi: 10.1002/ab.21601

Greenberg, J., Solomon, S., and Pyszczynski, T. (1997). Terror management theory of self-esteem and cultural worldviews: empirical assessments and conceptual refinements. Adv. Exp. Soc. Psychol. 29, 61–139. doi: 10.1016/S0065-2601(08)60016-7

Greenwald, A. G., and Farnham, S. D. (2000). Using the implicit association test to measure self-esteem and self-concept. J. Pers. Soc. Psychol. 79, 1022–1038. doi: 10.1037/0022-3514.79.6.1022

Gurtman, M. B. (1992). Trust, distrust, and interpersonal problems: a circumplex analysis. J. Pers. Soc. Psychol. 62, 989–1002. doi: 10.1037/0022-3514.62.6.989

Ha, J. H., Chin, B., Park, D. H., Ryu, S. H., and Yu, J. (2008). Characteristics of excessive cellular phone use in Korean adolescents. CyberPsychol. Behav. 11, 783–784. doi: 10.1089/cpb.2008.0096

Hankin, B. L., Lakdawalla, Z., Carter, I. L., Abela, J. R. Z., and Adams, P. (2007). Are neuroticism, cognitive vulnerabilities and self-esteem overlapping or distinct risks for depression? Evidence from exploratory and confirmatory factor analyses. J. Soc. Clin. Psychol. 26, 29–63. doi: 10.1521/jscp.2007.26.1.29

Haselhuhn, M. P., Kennedy, J. A., Kray, L. J., Van Zant, A. B., and Schweitzer, M. E. (2015). Gender differences in trust dynamics: women trust more than men following a trust violation. J. Exp. Soc. Psychol. 56, 104–109. doi: 10.1016/j.jesp.2014.09.007

Hawi, N. S., and Samaha, M. (2016). To excel or not to excel: strong evidence on the adverse effect of smartphone addiction on academic performance. Comput. Educ. 98, 81–89. doi: 10.1016/j.compedu.2016.03.007

Hayes, A. F. (2013). Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach. New York, NY: Guilford Press.

Ihm, J. (2018). Social implications of children’s smartphone addiction: the role of support networks and social engagement. J. Behav. Addict. 7, 473–481. doi: 10.1556/2006.7.2018.48

Isiklar, A., Sar, A. H., and Durmuscelebi, M. (2013). An investigation of the relationship between high-school students’ problematic mobile phone use and their self-esteem levels. Education 134, 9–14.

Jenaro, C., Flores, N., Gómez-Vela, M., González-Gil, F., and Caballo, C. (2007). Problematic internet and cell-phone use: psychological, behavioral, and health correlates. Addict. Res. Theory 15, 309–320. doi: 10.1080/16066350701350247

Jessor, R., Turbin, M. S., Costa, F. M., Dong, Q., Zhang, H., and Wang, C. (2003). Adolescent problem behavior in china and the United States: a cross-national study of psychosocial protective factors. J. Res. Adoles. 13, 329–360. doi: 10.1111/1532-7795.1303004

Jiang, Z., and Zhao, X. (2017). Brain behavioral systems, self-control and problematic mobile phone use: the moderating role of gender and history of use. Pers. Indiv. Differ. 106, 111–116. doi: 10.1016/j.paid.2016.10.036

Kavas, A. B. (2009). Self-esteem and health-risk behaviors among Turkish late adolescents. Adolescence 44, 187–198. doi: 10.1007/s11199-009-9600-1

Kim, E., and Koh, E. (2018). Avoidant attachment and smartphone addiction in college students: the mediating effects of anxiety and self-esteem. Comput. Hum. Behav. 84, 264–271. doi: 10.1016/j.chb.2018.02.037

Kim, J., Seo, M., and David, P. (2015). Alleviating depression only to become problematic mobile phone users: can face-to-face communication be the antidote? Comput. Hum. Behav. 51, 440–447. doi: 10.1016/j.chb.2015.05.030

Kim, S. S., Chung, Y., Perry, M. J., Kawachi, I., and Subramanian, S. V. (2012). Association between interpersonal trust, reciprocity, and depression in South Korea: a prospective analysis. PLoS One 7:e30602. doi: 10.1371/journal.pone.0030602

Kuss, D. J., Kanjo, E., Crook-Rumsey, M., Kibowski, F., Wang, G. Y., and Sumich, A. (2018). Problematic mobile phone use and addiction across generations: the roles of psychopathological symptoms and smartphone use. J. Technol. Behav. Sci. 3, 141–149. doi: 10.1007/s41347-017-0041-3

Lee, H., Ahn, H., Choi, S., and Choi, W. (2014). The SAMS: smartphone addiction management system and verification. J. Med. Syst. 38, 1–10. doi: 10.1007/s10916-013-0001-1

Lemola, S., Perkinson-Gloor, N., Brand, S., Dewald-Kaufmann, J. F., and Grob, A. (2015). Adolescents’ electronic media use at night, sleep disturbance, and depressive symptoms in the smartphone age. J. Youth Adoles. 44, 405–418. doi: 10.1007/s10964-014-0176-x

Lester, D., and Gatto, J. (1990). Interpersonal trust, depression, and suicidal ideation in teenagers. Psychol. Rep. 67:786. doi: 10.2466/pr0.1990.67.3.786

Lin, Y. H., Chang, L. R., Lee, Y. H., Tseng, H. W., Kuo, T. B., and Chen, S. H. (2014). Development and validation of the smartphone addiction inventory (SPAI). PLoS One 9:e98312. doi: 10.1371/journal.pone.0098312

Mackeen, B. A., and Herman, A. (1974). Effects of group counseling on self-esteem. J. Counsel. Psychol. 21, 210–214. doi: 10.1037/h0036491

Metalsky, G. I., Joiner, T. E., Hardin, T. S., and Abramson, L. Y. (1993). Depressive reactions to failure in a naturalistic setting: a test of the hopelessness and self-esteem theories of depression. J. Abnorm. Psychol. 102, 101–109. doi: 10.1037/0021-843x.102.1.101

Mortenson, S. T. (2009). Interpersonal trust and social skill in seeking social support among Chinese and Americans. Commun. Res. 36, 32–53. doi: 10.1177/0093650208326460

Murrell, S. A., Meeks, S., and Walker, J. (1991). Protective functions of health and self-esteem against depression in older adults facing illness or bereavement. Psychol. Aging 6, 352–360. doi: 10.1037/0882-7974.6.3.352

Orth, U., and Robins, R. W. (2013). Understanding the link between low self-esteem and depression. Curr. Direct. Psychol. Sci. 22, 455–460. doi: 10.1177/0963721413492763

Orth, U., Robins, R. W., Widaman, K. F., and Conger, R. D. (2014). Is low self-esteem a risk factor for depression? Findings from a longitudinal study of mexican-origin youth. Dev. Psychol. 50, 622–633. doi: 10.1037/a0033817

Panova, T., and Lleras, A. (2016). Avoidance or boredom: negative mental health outcomes associated with use of information and communication technologies depend on users’ motivations. Comput. Hum. Behav. 58, 249–258. doi: 10.1016/j.chb.2015.12.062

Robins, R. W., and Trzesniewski, K. H. (2005). Self-esteem development across the lifespan. Curr. Direct. Psychol. Sci. 14, 158–162. doi: 10.1111/j.0963-7214.2005.00353.x

Rodda, S., Brown, S. L., and Phillips, J. G. (2004). The relationship between anxiety, smoking, and gambling in electronic gaming machine players. J. Gambl. Stud. 20, 71–81. doi: 10.1023/b:jogs.0000016704.06088.85

Rosenberg, M. (1965). Society and the Adolescent Self-Image. Princeton, NJ: Princeton University Press.

Rotenberg, K. J., MacDonald, K. J., and King, E. V. (2004). The relationship between loneliness and interpersonal trust during middle childhood. J. Genet. Psychol. 165, 233–249. doi: 10.3200/GNTP.165.3.233-249

Rousseau, D. M., Sitkin, S. B., Burt, R. S., and Camerer, C. (1998). Not so different after all: a cross-discipline view of trust. Acad. Manag. Rev. 23, 393–404. doi: 10.5465/amr.1998.926617

Salehan, M., and Negahban, A. (2013). Social networking on smartphones: when mobile phones become addictive. Comput. Hum. Behav. 29, 2632–2639. doi: 10.1016/j.chb.2013.07.003

Samaha, M., and Hawi, N. S. (2016). Relationships among smartphone addiction, stress, academic performance, and satisfaction with life. Comput. Hum. Behav. 57, 321–325. doi: 10.1016/j.chb.2015.12.045

Simpson, J. A. (2007). “Foundations of interpersonal trust,” in Social Psychology: Handbook of Basic Principles , 2nd Edn, eds A. W. Kruglanski and E. Tory Higgins (New York, NY: Guilford Press), 587–607.

Smetaniuk, P. (2014). A preliminary investigation into the prevalence and prediction of problematic cell phone use. J. Behav. Addict. 3, 41–53. doi: 10.1556/JBA.3.2014.004

Söderqvist, F., Carlberg, M., and Hardell, L. (2008). Use of wireless telephones and self-reported health symptoms: a population-based study among swedish adolescents aged 15–19 years. Environ. Health 7:18. doi: 10.1186/1476-069X-7-18

Sowislo, J. F., and Orth, U. (2013). Does low self-esteem predict depression and anxiety? A meta-analysis of longitudinal studies. Psychol. Bull. 139, 213–240. doi: 10.1037/a0028931

Trzesniewski, K. H., Donnellan, M. B., Moffitt, T. E., Robins, R. W., Poulton, R., and Caspi, A. (2006). Low self-esteem during adolescence predicts poor health, criminal behavior, and limited economic prospects during adulthood. Dev. Psychol. 42, 381–390. doi: 10.1037/0012-1649.42.2.381

Wang, C., Lee, M. K. O., Yang, C., and Li, X. (2016). “Understanding problematic smartphone use and its characteristics: a perspective on behavioral addiction,” in Lecture Notes in Information Systems and Organisation: Transforming Healthcare Through Information Systems , Vol. 17, eds D. Vogel, X. Guo, H. Linger, C. Barry, M. Lang, and C. Schneider (Cham: Springer), 215–225. doi: 10.1007/978-3-319-30133-4_15

Wang, P., Zhao, M., Wang, X., Xie, X., Wang, Y., and Lei, L. (2017). Peer relationship and adolescent smartphone addiction: the mediating role of self-esteem and the moderating role of the need to belong. J. Behav. Addict. 6, 708–717. doi: 10.1556/2006.6.2017.079

We Are Social and Hootsuite, (2018). Digital in 2018: Essential Insight into Internet, Social Media, Mobile, and Ecommerce Use Around the World. Available at: https://digitalreport.wearesocial.com (accessed August 2, 2019).

West, S. G. (2011). Editorial: introduction to the special section on causal inference in cross sectional and longitudinal mediational models. Multivariate Behav. Res. 46, 812–815. doi: 10.1080/00273171.2011.606710

Wild, L. G., Flisher, A. J., Bhana, A., and Lombard, C. (2004). Substance abuse, suicidality, and self-esteem in South African adolescents. J. Drug Educ. 34, 1–17. doi: 10.2190/07c2-p41f-4u2p-jh0q

Xie, X., Dong, Y., and Wang, J. (2018). Sleep quality as a mediator of problematic smartphone use and clinical health symptoms. J. Behav. Addict. 7, 466–472. doi: 10.1556/2006.7.2018.40

Yamagishi, T. (1998). Trust and Social Intelligence: The Evolutionary Game of Mind and Society , trans. ed. T. Yamagishi (Tokyo: University of Tokyo Press).

Yamagishi, T., Akutsu, S., Cho, K., Inoue, Y., Li, Y., and Matsumoto, Y. (2015). Two-component model of general trust: predicting behavioral trust from attitudinal trust. Soc. Cogn. 33, 436–458. doi: 10.1521/soco.2015.33.5.436

Yamagishi, T., and Yamagishi, M. (1994). Trust and commitment in the United States and Japan. Motiv. Emot. 18, 129–166. doi: 10.1007/BF02249397

Yanagisawa, K., Masui, K., Furutani, K., Nomura, M., Ura, M., and Yoshida, H. (2011). Does higher general trust serve as a psychosocial buffer against social pain? An NIRS study of social exclusion. Soc. Neurosci. 6, 190–197. doi: 10.1080/17470919.2010.506139

Yang, Z., Asbury, K., and Griffiths, M. D. (2019). An exploration of problematic smartphone use among Chinese university students: associations with academic anxiety, academic procrastination, self-regulation and subjective wellbeing. Int. J. Ment. Health Addict. 117, 596–614. doi: 10.1007/s11469-018-9961-1

Yen, C. F., Tang, T. C., Yen, J. Y., Lin, H. C., Huang, C. F., Liu, S. C., et al. (2009). Symptoms of problematic cellular phone use, functional impairment and its association with depression among adolescents in southern taiwan. J. Adolesc. 32, 863–873. doi: 10.1016/j.adolescence.2008.10.006

Young, M. L., and Tseng, F. C. (2008). Interplay between physical and virtual settings for online interpersonal trust formation in knowledge-sharing practice. CyberPsychol. Behav. 11, 55–64. doi: 10.1089/cpb.2007.0019

Zung, W. W. K. (1965). Self-rating depression scale in an outpatient clinic. Arch. Gen. Psychiatry 13, 508–515. doi: 10.1001/archpsyc.1965.01730060026004

Keywords : problematic smartphone use, self-esteem, interpersonal trust, depression, moderated mediation model, adolescents, smartphone addiction

Citation: Li C, Liu D and Dong Y (2019) Self-Esteem and Problematic Smartphone Use Among Adolescents: A Moderated Mediation Model of Depression and Interpersonal Trust. Front. Psychol. 10:2872. doi: 10.3389/fpsyg.2019.02872

Received: 24 September 2019; Accepted: 04 December 2019; Published: 20 December 2019.

Reviewed by:

Copyright © 2019 Li, Liu and Dong. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Yan Dong, [email protected]

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.