Article summary with Reevaluating the strengths and weaknesses of self-report measures of subjective well-being by Lucas

What is this article about?

Well-being is subjective and because of that, all ways of evaluating well-being are subjective. The subjective nature of the construct makes self-report a natural method for assessing well-being. But there are problems with self-report measures. These problems will be reviewed in this article.

What is subjective well-being and how is it measured?

Subjective well-being (SWB) is a construct that focuses explicitly on subjective evaluations of one's quality of life as a whole. Different people may evaluate the same circumstances in different ways. The study of SWB relies on self-report measures. There is much debate on whether these measures are reliable and valid. There are some difficulties with self-report measures, but they are ultimately reliable and valid, as will be proven throughout this article. 

Many different self-report measures for evaluating SWB have been developed, but they all ask a person about the quality of one's life as a whole. Diener explained that there are affective components of how someone judges their life, based on how they feel, and cognitive components, based on how they cognitively reflect on their life. Traditionally, both the cognitive and affective components of SWB are assessed using global, retrospective measures that require respondents to think back on their lives and the characteristics of those lives.

Because of the potential for inaccurate reconstructions of one's affective experience earlier, some researchers believe that we should distinguish between measures that focus on retrospective evaluations and measures that capture people's emotional reaction to life as it is lived. This can be done with experience sampling methods, where respondents are signaled multiple times per day and asked to complete surveys about what they are doing, who they are with and how they are feeling. The downside of experience sampling methods, is that they are extremely time- and resource-intensive for both the researcher and the participants. This is why there are also mixed forms of methods.

Are subjective well-being measures reliable and valid?

There is no gold standard against which self-report measures of SWB can be compared. The evaluation of SWB is a subjective one and it captures internal thoughts and feelings that are not visible to outside observers. But this does not mean that SWB measures are unreliable or invalid. In general, SWB measures score well on quality of psychometric properties. Research has specifically shown that single-item measures of life satisfaction are mostly reliable. Reliability is one of the psychometric properties and is assessed with quantitative indexes. Validity is more complex and more difficult to establish in a definitive way. Validity is often defined as whether a test measures what it is supposed to measure, but it is difficult to prove that SWB is truly and totally measured. According to Diener et al. (2009) there are four types of validity that should be considered when evaluating self-report SWB measures:

• Face validity is the most simple form of validity and asks the question of whether a measure is asking the right questions. In this case: is the self-report SWB measure asking questions about SWB? Because SWB is so broad, SWB measures usually have face validity.
• Content validity: does a measure capture the breadth of the construct of interest, without including content that should be excluded. This is more difficult for SWB because there are a lot of things that might be part of SWB, but that we are not sure of yet. "Do you think you look attractive" might be connected to the question of SWB, but we are not sure what that connection is yet.
• Convergent and discriminant validity. Convergent validity reflects the extent to which a measure correlates strongly with other related measures. It is important to show that different measures of the same construct cohere even when assessed using different methods of assessment. Discriminant validity reflects the extent to which a measure correlates weakly or not at all with measures to which it should be unrelated. SWB measures tend to show reasonable levels of convergent and discriminant validity. 
• Construct validity is the most complicated form of validity to assess. Construct validity reflects the extent to which a measure behaves as it would be expected to behave, given theories about the construct itself. Because SWB is so broad and we still have a lot to learn about SWB, we don't always know what to expect. But this is a difficult form of validity for all measures and especially for measures that measure big concepts that have only recently been researched. 

What is the judgment model of subjective well-being?

A big challenge to the validity of SWB measures comes from the judgment model of SWB by Schwarz and Strack (1999). This model starts with the assumption that when asked about your well=being, you will not have a response stored in memory that can simply be accessed and reported. Instead, you will have to construct a response at the time of judgment. You will construct this on the basis of what domains of your life seem relevant at that moment and for that question, because you do not have the time or the capacity to consider all relevant domains of life. 

Because of this model, broad concerns have been raised about standard research practices in the field, especially within specific domains in social psychology. There is a lot of methodological critique that SWB measures can lead to high rates of false positive findings, because you only think about something because you are asked about it, but it does not actually play a big role in your life. When asked about your romantic relationship for instance, you will think about the last argument you had with your partner, even though your relationship is otherwise fine. You are simply not able to oversee the bigger picture and automatically think about the last big event. 

This is indeed a cause for concern, but researchers can keep this in mind by asking the right questions and taking the model in account in their conclusions. 

Existing research thus suggests that SWB measures typically have desirable psychometric properties, including relatively high levels of reliability and convergent, discriminant and construct validity. But SWB measures are not perfect. This is why researchers have to take the judgment model of SWB in account. Research on the properties of self-report measures can not only strengthen conclusions from research that uses those kinds of methods, but can also help clarify what SWB is and how people evaluate their lives.