Scoring Instructions for the DSES (updated January 2022)
Refer to the www.dsescale.org home page on the scale before using it, and also register with lynn@lynnunderwood.com.
When using the DSES quantitatively, items 1-15 are scored: Never 1, Once in a while 2, Some days 3, Most days 4, Every day 5, Many times a day 6. Item 16 is scored: Not at all 1, Somewhat close 3, Very close 5, As close as possible 6.
Mean score
The best way to score the whole DSES is using a mean score – adding all item scores up and dividing by 16. This is better than a total sum score for the following reasons: (1) using mean scores allows for better examination of individual items or subgroups, (2) it can compensate for missing responses, as you can divide the total by the number of answered questions to get a mean score, and (3) mean scores also more easily allow comparison with other research that used a 6-item version or those using only 15 items.
Interpreting group scores
Average scores for a group of people are very useful for comparing one group with another, or looking at how DSES scores correlate with other variables examined in particular studies. There are no set high, medium, or low scores for the DSES that can be applied universally. Retaining the group mean scores as continuous variables, rather than grouping data into high, medium, and low, better allows for comparison across studies. In fact, for many other scales the grouping into these categories is an ad hoc division that is not particularly linked with practical outcomes or correlations. Regarding norms, the DSES was put on the General Social Survey in the US (Underwood 2006), which established norms for a random sample of the US population. Other individual country-level studies, and other subgroup studies, have established norms that can be useful for comparison of populations. Finally, within each research study, investigators can find natural breakpoints and distribution within their data to guide interpretation based on prior hypotheses of interest.
Interpreting individual scores
Number scores are not appropriate for comparing specific people with one another. To look at one person’s score and compare it with another’s can be counterproductive. People start from different cultural backgrounds and substrates of emotional sensitivity and sets of beliefs. This evens out when looking at whole group averages but can obscure things when comparing one individual to another. In individual assessment in practical application contexts it is particularly important that there are no absolute high, medium, or low levels of ordinary spiritual experience. Each individual has specific personality and sensitivity variables, as well as different life situations and levels of comfort with religious language. It doesn’t work to compare individual scores with one another in a treatment or organizational setting. However, it can be quite useful to look at changes in a person’s own score over time for a particular person.
Change over time.
A valuable use of the DSES in its quantitative form is to examine how scores vary over time, whether following an intervention such as a treatment or educational regime, a discrete set of events, a long-term life experience, or a developmental trajectory. After some change or intervention, or merely to follow the experiences and resources of a client or participant, you can compare the scores of a single person over time using the original version or the checklist version. Likewise, you can use group averages to look at correlations with other factors and outcomes over time to answer questions about causation and interactions of variables. In a smartphone study, Wright and colleagues (2017) calculated an overall mean for the person (trait level DSES) and then examined deviations from the person’s mean at each point that data was collected, comparing where the person was relative to their average. That is, they measured the variation in DSES above and below normal for that person. The variations around their mean were predictive of certain outcomes such as less feelings of stress, or more capacity to have a caring attitude for those around them or strangers.
Subgroup analyses.
Most studies have shown a unifactorial loading of all the 16 items on a single factor, and high inter-item correlation using statistical methods such as Cronbach’s alpha (Underwood 2002, 2011). This encourages us in using a mean score for the whole DSES in research and practice. That has been the case in most of the studies reviewed. Psychometric properties for translated versions of the DSES are also strong and usually unifactorial, with Cronbach’s alphas ranging from 0.86 to 0.98.
However, some studies have shown two factors. Especially in populations with a substantive group of people for whom the word “God” may not be comfortable, there can be a tendency for the items containing that word to be reported less frequently than others, or to be differentially correlated with outcomes. It is important to note that some studies finding such a division have used the scale without the introductory sentences that allow for substitution language for the word “God.” Inclusion of the introductory sentences and offering formats with alternative in-line language help soften this division, but it may still remain. In the original DSES format, eight of the items mention the word “God,” and eight do not. If separated into two groups in analysis, the researchers often call them theistic and non-theistic, or theistic and self-transcendent. These labels can be misleading, as many who are theistic score highly on the ‘non-theistic’ items, and many of the theistic items represent ‘self-transcendence.’ These labels assigned to subgroups can unduly influence interpretations of results and implications for action. For example, “I feel a connection to all of life” can be particularly salient for non-theists. However, as environmental awareness increases, this may also be increasingly relevant for theists in many Christian, Jewish, Muslim, Hindu and indigenous theistic traditions. And even those who categorize themselves as atheists can express implicit attitudes toward a transcendent reality when they report experiences of some of the items containing the word “God,” especially when using the substitution language in the introductory statements or inline alternatives. An example of this complexity was demonstrated in a study looking at people in the Basque area of Europe, where they found that atheists actually scored higher on the full DSES than those self-identifying as Christian (Mayoral Sánchez, Laca Arocena, and Mejía Ceballos 2010). The DSES was designed based on research demonstrating that some people may not identify as religious, yet use much religious language, and even continue many religious practices, or use religious frameworks in their approach to life. Explicit beliefs and implicit attitudes are not necessarily linked. Self-reported belief or unbelief is often belied by attitudes and feelings, and cognitive adherence to a belief set may not map onto some factors of significance. The researcher or practitioner should not create boundaries or categories that may not exist in the mind of the respondent. Cultures will also differ as to if, or how, this kind of division might be useful. We would expect that as the use of the scale evolves and increasing use of alternative inline language or ability to actually substitute one’s own word choice becomes more technically feasible, results will tend even more towards a unifactorial solution.
Another subgroup of items that has been used are the two items: “I feel a selfless caring for others” and “I accept others even when they do things I think are wrong.” These items have been labeled the compassionate love items and analyzed separately, especially in studies of relationship quality and prosocial behaviors (Underwood 2009).
Individual items
A few studies have either used only individual items from the DSES or analyzed individual items separately. Each individual item in the DSES taps one particular aspect of ordinary spiritual experience, and particular items may be especially salient for particular research hypotheses or useful personally or professionally. The deep inner peace or harmony item, the thankful for blessings item have been used in studies as single items. The items that tap into a sense of transcendent compassionate love may also, for example, be ones worth focusing on in particular settings or research.
Checklist version scoring
The introductory words delimit the time frame: “Recently” can also be delimited more discretely, as in ‘Today,’ ‘In the last hour,’ etc.)
The response categories for all 16 items are scored: No 0 Yes 1
For link to a copy of the checklist version see here.