How to create frequency variables in SPSS (PAWS) when you use an Experience Sampling design

This post is mainly for people who use ESM protocols and who want to answer research questions related to frequencies in activities, social contexts, etc, such as:

1. Are individuals who spend more of their time in leisure activities happier?

2. Is spending more time with family during daily life associated with higher positive affect when alone?

3. Is the amount of time spent in active leisure during everyday life associated with personality traits?

The questions are merely posed to explain different rationale and possibilities when using frequencies to assess the amount of time individuals reported in different contexts during the ESM assessment.

For those who do not know the Experience Sampling Method, this method aims to assess individuals’ experiences as they go about their daily life. It captures moments of individual’s days and weeks and assesses physical and social contexts, activities, thoughts and subjective experience. Participants are asked to carry either electronic beepers or wristwatches that randomly emit acoustic signals throughout the day. Participants fill in a questionnaire every time they hear the signal. The common duration of a ESM protocol is one week. However, the aims of the study may require longer or periodic assessment periods and interval-contingent or event-contingent reporting.

Researchers who use this method end up with a gigantic, never-ending dataset that has information about each variable in each report, of every day, week and sampling period, for each individual. For example, if participants filled in the 7 ESM reports per day, during 7 days (assuming participants were 100% compliant), the dataset would comprise 4900 lines in a study with 100 participants.

However, questions that relate to person-level characteristics need to use variables created based on all the samples for each subject. For example, the aim of the study might be to assess overall levels of happiness. In this case, we aggregate all the scores for happiness (and other items, if adequate) reported by each individual during the ESM assessment period. This will give us one single value of “happiness” per subject. This is easily done using SPSS:

Data >> Aggregate

1. Set “subject” as the Break variable

2. Set “happiness” as the Aggregated variable and choose the Function of Mean (it usually is the default)

Now, you might also want to investigate whether spending time in a context, activity, construct (e.g., flow) is related to other outcomes. In the examples I gave in the beginning of this post, I had three questions that can be analyzed at different levels, using these frequencies. Remember that these frequencies refer to person-level variables, thus, reflecting between-individual differences, as opposed to within-individual changes (beep-level).

In the first question (1) There are two person-level variables derived from aggregated data: relative frequency in leisure, and mean happiness; in the second question, (2) there is one person-level variable (frequency of time spent with family) and two beep-level variables (momentary positive affect and social context); Finally, (3) the third question relates to the association between a person-level variable derived from aggregating data from the ESM (time spent in leisure activities) and information gathered from other questionnaires (e.g., NEO-PI, STAI, etc.).

Some of you may know how to do these frequencies, while others may not, and it is because of the latter, and because I went to too much trouble until I found out how to do it using PAWS (aka, SPSS) that I’m leaving it here.

The process is the same as the one used to aggregate the individual’s data to create individual scores based on means or standard deviations:

Data >> Aggregate

1. Set “subject” as the Break variable

2. Set the categorical variable you wish to analyze as the Aggregated variable (e.g. Social Context). ***Be sure you know the values of your variable!***

3. Go to the Function menu.

4. In Percentages, select the option that better fits your needs:

4.1. If you have a dichotomous variable (0/1) you can choose above 0 to get the percentual frequency individuals reported being in the 1 context; or below 0 if you want to know the percentual frequency in which individuals reported being in the 0 context during the assessment week.

4.2. If you have a categorical variable with different values (e.g., 0-4) you can choose inside or outside to assess the value(s) you want to analyze.

Although researchers only have a week of assessment for each individual, this week is usually chosen to be a representative week of everyday life. Also, assessing time spent in different contexts based on ESM reports seems to be congruent with what individuals report doing using other diary methods. This indicates that this is a good way to evaluate individuals’ time budgets.


One thought on “How to create frequency variables in SPSS (PAWS) when you use an Experience Sampling design

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s