Parenting is difficult to measure, particularly in cultural context and with developmental sensitivity. These are three measures I have used in my own studies to assess dimensions of parenting style.

Parenting Style Inventory II

I began developing the Parenting Style Inventory (PSI) in 1993, after completing an extensive review of the historical literature on parenting style that introduced the distinction between parenting style and parenting practice (Darling & Steinberg, 1993).

Darling, N., & Steinberg, L. (1993). Parenting style as context: An integrative model. Psychological Bulletin, 113(3), 487–496. https://doi.org/10.1037/0033-2909.113.3.487

The goal of the PSI-II measure was to introduce a relatively brief measure that assessed adolescents’ perception of their parents’ demandingness, responsiveness, and psychological autonomy-granting independent of the specific parenting practices. The measure was first developed for use with middle school students in Pennsylvania. I have since used it in studies of 12-22 year old students in the US (Central PA, OH, and Miami, FL), Santiago, Chile, and Manila, Philippines.

This measure has been widely used around the world and has been translated into several local dialects of Spanish as well as Tagalog. I have been informed that it has been used successfully across the globe.

There are several issues researchers should think about when they choose to use this or any other measure of parenting.

• All measures of parenting style tend to have strong measures of parental warmth and support. That has been true for the PSI-II as well.
• The construct we call ‘demandingness’ and others refer to as ‘control’ is more complex. Virtually all measures have difficulty getting strong internal consistency in all contexts. Our measure is designed to assess whether parents expects their child to follow the rules that they set and follows through if they do not. It is not a measure of strictness per se – a parent could set many rules and not enforce them or set few rules and be very firm about them. It is about contingency.
• Internal consistency relies on having a good distribution in your variables. If parents in a particular setting are ALL very high in enforcement, the demandingness measure will show low internal consistency. Check the measure in your context to see if it works.
• Negatively worded items can be confusing. For example, a double negative in English is positive. A double negative in Spanish is negative. You may want to reword negatively worded items when working with young students, poor readers, or countries where the meaning will be unclear. Test! Test! Test!

PSI-III

To deal with some of the complexities of negatively worded items and to better address the issue of measuring demandingness, we began working on the PSI-III. This is what I currently use in my own work.

As you can see, looking at a comparison of items from the PSI-II and III, it is still in development. The demandingness subscale has more items than I use in order to test the items in different contexts. If one doesn’t work, we drop it. What I like about this subscale is that it taps what I think Baumrind is getting at in her original work on parenting style. The core items (1, 2, 5, 6, 7, 8, and 11) all measure parental expectation that the child act responsibly and maintain control of themselves. In her late work, she talks about how authoritative parents balance accomodating the needs of the child (support) with the requirement that they child modulate their behavior in consideration of others. That is what demandingness here is trying to capture: the expectation of self-regulation.

Note that this measure is specifically not about strictness, although there are items about monitoring (9) and enforcement (3). Two experimental items ask about being respectful of others – again trying to capture self-regulation. This measure is about expecting self-regulation rather than exertion of external power. I measure emotional control using Barber’s measure of that construct. The core items in this measure perform well in all the contexts I have used them.

Using the Parenting Style Inventory II or III

I am often asked how to score the PSI-II and III.

As described in Darling & Toyokawa (1997), items for each subscale are reverse coded and a mean is calculated. I use a mean rather than a sum so that missing items do not artificial lower parent scores.

I would note that how you deal with multiple parents in the household when evaluating parenting influences is a complex decision that needs to be carefully considered. Adolescents often report on only one parent – usually the mother. Eliminating those without fathers is certainly problematic in terms of representing a population. One option we have sometimes used is to use the score of the parent that is highest – in other words, their most functional parent. This has issues as well. If using a categorical classification of style rather than a linear measure of dimensions, one can categorize adolescents as having different constellations of parents.

A second major decision is whether to use the subscales as linear measures or to use them to create categorical classifications of parenting styles (e.g., authoritative, permissive, etc.).

To create a categorical classification, many people do a median split of support and demandingness, categorizing adolescents as to where they fall relative to other families in the sample.

Alternatively, I have used a K-means cluster analysis to create ideal types (high support, demandingness, and autonomy-granting; high support, low demandingness, etc.) that correspond to Baumrind’s original descriptions. That has the advantage of being conceptually clear. More importantly, each participant’s fit to the archetype can be computed, and those who do not fit the category well can be eliminated from the model. This can be advantageous because in normally distributed samples, most people are in the middle and, although classified as different family types, are actually quite similar to one another.

Finally, we have sometimes computed a single ‘authoritativeness’ score that simply sums all 15 core positive items together. This can be useful in small samples where style is being used primarily as a control or moderator variable rather than the core focus of the study. Although she raised her eyebrows at this technique, Baumrind did describe it as an ‘intriguing approach’.

Escale Parental Brave

The Escale Parental Brave is a brief parenting style measure developed from the PSI-II that more systematically addresses its limitations. It has much stronger psychometric properties and was validated in both Chile and Italy. I include both the published paper and a poster explaining it’s development and properties.

Cumsille, Martinez, Rodriguez, Darling (2011)