Critique of the Data Analysis: Kraut/kiesler

Critique of the Data Analysis: Kraut/Kiesler study.

Though it has been many years since the Kraut/Kiesler study was conducted, the spirit of the study still holds interest to contemporary behavioral research if for nothing more than as a gauge by which to understand the evolution and change in user depth of or reliance on the internet, as it might become for one, a crutch for the semblance of personal interaction. Kraut et al. (1998) conducted a longitudinal study at the seminal incorporation of home internet use into the lives of respondents (p. 1017). Longitudinal studies allow for better data set results. These types of studies are not hindered by biases of respondents at a moment in time or by extenuating variables not able to be tested or not focused on by researchers. The authors posit in this pre-Facebook time period, "that internet use could have enormous consequences for society and for people's personal well-being" (p. 1017). The authors attempt to present some of those 'consequences' by testing their model based on theory associated with Social Impact Analysis.

Building the Model

The authors conducted a longitudinal study to examine causal relationships between people's use of the internet, their social involvement, and certain likely psychological consequences of social involvement (p. 1020). The approach taken for building the model for this study was based on based on path analysis. The aim of path analysis is to provide estimates of the magnitude and significance of hypothesised causal connections between sets of variables (Webley and Lea, 1997). Path analysis was used to test relationships among variables measured at three time periods: pretest questionnaire at Time 1 (T1), Internet usage during Time 2 (T2), and posttest questionnaire at Time 3 (T3) to build the model for this study based on 'Logic of Social Impact Analyses' (p. 1023). The T paths are laid out in Figure 1 below.

Figure 1. Logic of Social Impact Analyses (1998)

These time periods are well established and present the authors with excellent reference points by which to gauge the changes of respondents over the study as designed. The data collection revolved around demographic characteristics to account for social involvement status. Not so much with regards to social standing but more leans towards levels of social interactions per respondent. A questionnaire was developed to collect variable data for the model which included demographic variables (age, gender, and race) used as control variables, internet usage, personal electronic mail use, World Wide Web use, and social involvement and psychological well-being (p. 1021).

All of the data collected was primary data. There was no use of archival data, focus groups or interviews. This allows for the