Theme 4: Quantitative research

This week I decided to begin by reading the paper Physical activity, stress, and self-reported upper respiratory tract infection by Fondell et al. Hereafter I will refer to upper respiratory tract infection as URTI. As we see in the paper quantitative methods like online surveys can be an extremely effective tool, as it allows researchers to draw conclusions from very large amounts of data. Using qualitative methods it is almost impossible to reach generalized conclusions that can be applied to the whole population or a larger group. However, in the paper some of the weaknesses of quantitative methods also become apparent. For example, there is no way for the researchers to verify the information submitted is correct. In the conclusions we don’t see a statement about URTI, but rather about self-reported URTI. Also, the researchers are unable to find more detailed information that the respondents can’t be expected to know themselves, in this instance the researchers couldn’t know for sure if a respondent had influenza or a common cold.

Keeping this in mind I went to read the research paper I chose: Measuring Mobile Phone Use: Self-Report Versus Log Data, by J. Boase and R. Ling. It was published in Journal of Computer-Mediated Communication with an impact factor of 1.778. I was happily surprised that the paper attempts to answer some of the questions I had when reading the paper on URTI, namely the reliability of larger surveys. The aim of the study is to compare self-reported data about mobile phone use to server logs supplied by a network operator. The quantitative methods used are an internet based survey done in Norway in 2008 and analysis of the phone log data of 613 respondents who authorized it when answering the survey. The researchers then compare the responses of the survey to the average phone usage pulled from the logs and calculates the correlation between the numbers. The advantage of using quantitative methods in this instance, apart from the researchers pronounced objective to evaluate and compare quantitative methods, is that the the researchers are able to draw much broader conclusions than they would if they had done a qualitative study on a smaller sample group.

I found the results of the study very interesting, as the researchers conclude that the self-reported data only correlate moderately with server log data. They also found that asking respondents to estimate “how often” they use their phones gave more accurate responses than asking how much they used their phone “yesterday”. The yesterday-question was something I reflected on while first reading it. My first thought was that it would not produce very good results, as, at least for me, the amount I use my phone varies heavily from day to day. Since it would be compared to the monthly average pulled from phone logs I figured there was bound to be deviations. On the other hand, I would have a hard time correctly estimating how much I use my phone on average so I found it interesting that that question still produced the best results. What I take away from this study is that what questions you ask, especially when it comes to surveys and self-reporting, heavily affects the quality and accuracy of your data.