Do young people prefer older psychotherapists?

Abstract

In times of demographic change, psychotherapists tend to stay longer in their jobs. Against the background of two contradictory age stereotypes (wise/generous versus senile/outdated old person), this analogue study investigates young adults’ preference for old over young psychotherapists, depending on presenting problem. In a within-subjects design, therapy-motivated young female participants (N = 79) received two kinds of hypothetical presenting problems, ‘universal problems’ (addressing fundamental questions of life virulent across the adult life span) and ‘young problems’ (life events that are developmentally close to and specific for young adulthood in today’s world). For each presenting problem, participants were presented with two naturalistic photographs of an old (55 + years) and a young (< 35 years) female psychotherapist; in a two-alternative forced choice decision task, participants were then asked to decide whom of the two psychotherapists they would prefer. Participants had neither an overall preference for old nor for young psychotherapists. As expected, we found a preference for old psychotherapists in case of ‘universal problems’ (friend’s death, natural disaster, quarreling with one’s past). In turn, a preference for young psychotherapists was observed in case of ‘young problems’ (cyberbullying, paternalizing parents, lovesickness). Effect size was medium-to-large. Organizations that provide services to younger people may want to inquire about a patient’s preference about age of a therapist prior to treatment. As catering to patient’s preferences for therapy has been shown to lead to better engagement and therapy outcomes, health care providers should try to meet the patient’s needs.

Introduction

Chronological age has been shown to have no validity for predicting job performance across different jobs (Schmidt et al. 2016). This also seems to hold true for psychotherapists (Miller et al. 2013). However, older workers are often confronted with age stereotypes of being less flexible, productive or ‘up-to-date’ (Posthuma and Campion 2009). Studies consistently showed that presenting one’s chronological age in a resume may lower older applicants’ job chances, at least in case of young recruiters (Richardson et al. 2013). Similarly, older persons are usually rated more negatively on evaluative statements in the domain of work (Kornadt and Rothermund 2011). In line with the positive age stereotype of the wise old person (Kite et al. 2005), one may argue, however, that these negative age stereotypes do not apply for selection of psychotherapists. Accordingly, younger patients might prefer older psychotherapists, as the latter are perceived as more generous and better able to solve problems against the background of more professional and life experience.

Our analogue study investigates therapy-motivated young adults’ preference for psychotherapists based on age (old age vs. young age of psychotherapist). In times of demographic change, psychotherapists tend to stay longer in their jobs (Gesundheitsberichterstattung des Bundes 2017). While various studies have investigated gender as one factor that people consider when choosing a mental health provider (Landes et al. 2013), very little is known about how age cues influence young adults’ initial selection of a psychotherapist. In the face of the high prevalence of mental disorders in the twenties, there is the need to make adequate treatment easily available for young adults in order to help them master their developmental tasks and avoid mental disorders in consecutive developmental periods (Gustavson et al. 2018).

Our research question and design build on research on age stereotypes and evaluations of older people. Accordingly, advanced age is readily visible in physical appearance (particularly by wrinkles) and automatically evokes judgements about a person’s abilities and competencies (Nelson 2002). Our study is further qualified by the large number of findings showing that contextual factors have a great impact on the activation of age stereotypes (Casper et al. 2011). For example, negatively biased evaluations of older workers in hiring decisions have been shown to be only apparent in contexts in which specific attributes (e.g., ‘dynamic company’) were made salient that contrasted with work-related stereotypes of old people (Diekman and Hirnisey 2007). In line with such a contextualist framework of age stereotypes, we examined differences in therapy-motivated young people’s preference for hypothetical psychotherapists of different ages (old versus young) under varying contextual therapeutic conditions, namely two different types of hypothetical presenting problems.

Specifically, we assumed that young people would prefer older psychotherapists in treatment occasions dealing with life problems addressing fundamental questions of life virulent across the adult life span, including strokes of fate and loss experience (‘universal problems’). In other words, for these presenting problems, we assumed that young adults would prefer older psychotherapists, given their perceived superiority in life experience and insight into the human condition. At the same time, we hypothesized that a preference for old psychotherapists would not be observed or even turned into a preference for young psychotherapists in case of treatment occasions involving critical life events that are developmentally close to and specific for young adulthood in today’s world (‘young problems’), such as peer relationship problems and virtual world problems. In our study, we kept female gender constant across participants and hypothetical psychotherapists in order to avoid confounding effects. Female participants were chosen as women are more likely than men to seek psychotherapy in case of mental health problems (e.g., Call and Shafer 2018). Female hypothetical psychotherapists were chosen given the distinct preponderance of women among psychotherapists (about 75 percent in Germany, German Federal Chamber of Psychotherapists 2018).

Method

Participants

The study was announced as a survey on ‘general opinions and attitudes towards psychotherapy.’ Female participants, aged between 18 and 29 years, were recruited via the online platform Unipark in Germany. From the initially recruited 105 participants, 13 terminated their participation before completion (response rate of 85.1%). Thirteen more participants were excluded as they indicated not to be motivated to see a psychotherapist in case of a psychological problem. The final sample comprised 79 participants who indicated that they were motivated to see a psychotherapist in case of a psychological problem (i.e., analogue sample). The mean age was M = 23.15 years (SD = 3.47), of whom 70.1% had a German Abitur. Twenty-one participants (26.6%) had therapy experience, five participants with an old psychotherapist (55 + years). (In work contexts, ‘older workers’ usually refers to people aged 55 plus, OECD 2018). An a priori power analysis indicated that a total sample size of N = 58 was necessary for detecting a small-to-medium effect (d = .40), with 90% power using Wilcoxon signed-rank test with α = .05, which we deemed the minimal sample size in order to test our hypotheses.

Procedure and design

Prior to the main study, study materials (i.e., hypothetical presenting problems and photographs of psychotherapists) were developed and a series of pilot studies were conducted to ensure relevance and equivalence of study materials, see ‘Appendix.’

Using a within-subjects design, each participant received the same six fictitious scenarios describing various critical life situations that may represent presenting problems in counseling and psychotherapy. The order of assignment was random in order to avoid sequence effects. Three of the six hypothetical presenting problems (i.e., 50%) represented ‘universal problems’ (friend’s death, natural disaster, quarreling with one’s past) or ‘young problems’ (cyberbullying, paternalizing parents, lovesickness), respectively. For each presenting problem, participants were asked to vividly image that they would find themselves in this situation. In a two-alternative forced choice decision task with two side-by-sided naturalistic photographs of an old versus a young female psychotherapist (55 + years versus < 35 years), participants were asked to decide whom of the two psychotherapists they would prefer if they were considering to see a psychotherapist for this problem. Photographs of psychotherapists (six old-young pairs in total) rather than descriptions were chosen as they make recipients feeling emotionally closer to a situation (Winkielman and Gogolushko 2018), and as they simulate first contacts between patients and psychotherapists in real life.

Statistical analyses

Given the dependency of the two choices (each participant made choices with respect to both kinds of presenting problems), hypotheses were tested using the Wilcoxon signed-rank test. To do so, two variables were computed, one reflecting the sum of choices concerning the ‘young problems’ and the other one reflecting the sum of choices with respect to the ‘universal problems.’ Since the choices were coded as 0 (choice for young psychotherapist) or 1 (choice for old psychotherapist), higher sums indicate a preference for the old psychotherapist.

In order to back up these results and specify the effects, we additionally ran logistic regression analysis to test our hypotheses (e.g., Cohen et al. 2003; Everitt 2001). In logistic regression, the logit of choosing one of two options (e.g., choosing either a young or an old psychotherapist) is assumed to be a linear function of the independent variables, as is depicted in Eq. (1):

$$\text{logit}\left(p\left(\text{old}\right)\right)=ln\left(\frac{p\left(\text{old}\right)}{1-p\left(\text{old}\right)}\right)=B_0+B_1{X}_1+\cdots +B_n{X}_n$$ 1

where p (old) is the probability of choosing an old psychotherapist, B₀ is a constant (the intercept), x₁ to x_n are predictors assumed to affect that probability, and B₁ to B_n are the regression weights. In standard logistic regression, it is assumed that all observations in the design are independent. Since our participants made six choices in succession, making all observations more or less correlated within participants, we used the generalized estimating equation (GEE) approach (Liang and Zeger 1986). In this approach, observations are assumed to be dependent within subjects and independent between subjects. The predictor in the regression model was the kind of presenting problem (‘young problem’ vs. ‘universal problem’), and the binary dependent variable was the participants’ choice (young vs. old psychotherapist).

Results

Table 1 shows how the 474 choices made by the N = 79 participants (six choices per participant) are distributed by presenting problem. Table 2 shows the participants’ choices favoring the old psychotherapist, related to the type of presenting problem. The mean of the sum of choices for the ‘young problems’ was M = 1.01 (SD = 0.90), and for the ‘universal problems’ M = 1.80 (SD = 1.07). Wilcoxon signed-rank test produced a significant difference between both values, z = − 4.28, p < .001. The z value can be used to calculate an effect size, such as the r proposed by Cohen et al. (2003), resulting in r = .34. In order to facilitate the comparison with other studies, we converted r to Cohen’s d (Borenstein et al. 2009), yielding d = .72 which corresponds to a medium-to-large effect. This result shows that participants preferred old psychotherapists over young psychotherapists in case of ‘universal problems,’ whereas the reverse was true in case of ‘young problem.’

Table 1.

Absolute frequencies of participants’ choices by presenting problem

Treatment occasion	Choices		Sum
	Young therapist	Old therapist
Young problem	157	80	237
Universal problem	95	142	237
Sum	252	222	474

The number of participants was N = 79, each of whom made six choices

Table 2.

Choices favoring the old psychotherapist by presenting problem

Treatment occasion	Choices in favor of the old therapist
	Sum	Mean	SD
Young problem	80	1.01	0.90
Universal problem	142	1.80	1.07
Sum	222	1.80	1.07

Choices favoring the old psychotherapist were coded as 1, choices favoring the young psychotherapist were coded as 0

As depicted in Table 3, logistic regression analysis produced a significant effect of the kind of presenting problem. Specifically, the logit increases by B = 1.08 when the participants were presented with a ‘universal problem.’ This effect translates to an odds ratio of 2.93, which indicates that the odds of selecting an old psychotherapist is increased by a factor of 2.93 when the participants were presented with a ‘universal problem’ instead of a ‘young problem.’ Controlling for participants’ age and ‘therapy experience with an old psychotherapist’ did not change the magnitude of the results.

Table 3.

Results of logistic regression analysis (N = 79 participants)

Parameter	B (SE)	Exp (B)	Wald χ2 (df)	p
Intercept	− 0.67 (0.15)	0.51	20.27 (1)	< .001
Kind of problem	1.08 (0.23)	2.93	22.13 (1)	< .001
QIC	627.74

QIC means quasi-likelihood under the independence criterion. Coding of the binary predictor was as follows: 0 = ‘young problem’; 1 = ‘universal problem.’ The binary variable was coded as 0 (choice of young therapist) or 1 (choice of old therapist)

Discussion

While the role gender has been a topic of research in various studies on treatment preferences, our analogue study is the first to investigate how age influences the initial selection of a psychotherapist. Overall, we found that therapy-motivated younger women had neither an overall preference for old nor young female psychotherapists based on naturalistic photographs. As expected, we found ‘an older psychotherapist favoritism’ in case of presenting problems dealing with universal life problems virulent across the adult life span such as strokes of fate and loss experiences. In turn, ‘a younger psychotherapist favoritism’ was observed in case of presenting problems involving critical life events that are developmentally close to and specific for young adulthood in today’s world, such as peer relationship problems and virtual world problems.

Our results speak for an age-matched effect in young people’s preference for an old over a young psychotherapist depending on presenting problem. Our findings are in line with a contextualist framework according to which age is associated with different abilities and competencies depending on circumstances such as the domain of evaluation (Casper et al. 2011; Kornadt and Rothermund 2011). In case of ‘universal problems,’ the positive age stereotype of the wise, life-experienced and generous old person may be readily activated. In the case of ‘young problems,’ old age may be associated with obsolescence, being outdated and psychological distance from the problems virulent on young adulthood in today’s world; alternatively, young psychotherapists might be preferred as generational peers ‘at eye level’ from which one feels understood without much explanation.

Our study does not only add to literature on the effects of age stereotypes inside and outside of work contexts (Weber et al. in press), but also extends the body of literature on the impact of psychotherapist characteristics (e.g., gender, ethnicity) that people consider when choosing a mental health provider (Eells et al. 1999).

Limitations and future research

Rather than conducting an analogue study, it would have been preferable to survey with a clinical sample of actual patients seeking therapy. Study results may have been more valid had participants been currently experiencing the problem they were asked to consider. The generalizability of the findings is limited to young, rather well-educated women and their preference for female psychotherapists. Investigating the effects of interactions between psychotherapists’ gender/age and patients’ gender/age would be important, by drawing on multifactorial designs. Furthermore, future research should investigate preferred age of psychotherapists in people of various ages including children, adolescents, middle-aged and older people. In order to bring light into the underlying causes of the low demand for psychotherapy among older adults, future research should test whether older people have a bias against younger psychotherapists and rather prefer psychotherapists of the same age.

Working within an age stereotype paradigm, we did only assess evaluations of old psychotherapists and the ‘contrasting’ group of young psychotherapists; investigating judgements of middle-aged psychotherapists should be investigated. Furthermore, future studies should investigate how experiences with older psychotherapists inside and outside of therapy moderate people’s initial preferences for old psychotherapists. In addition, further research is needed to investigate how explicit and implicit beliefs about old age influence people’s preference of old versus young therapists. Future research should also investigate how (old) age of psychotherapists influences other outcomes such as anticipated comfort self-disclosing. In order to disentangle the effects of age and level of professional qualification, future research should systematically vary both dimensions in descriptions of psychotherapists.

Implications

Organizations that provide services to younger people may want to inquire about a patient’s preference about age of psychotherapist prior to treatment, even though most patients do not state a preference regarding characteristics about a psychotherapist (most likely the gender, Speight and Vera 2005). As catering to patient’s preferences for therapy has been shown to lead to better engagement and outcomes (Swift et al. 2011), health care providers should try to meet the patient’s needs. Findings may be particularly relevant in cases where time is limited, either in short-term therapy or crisis services. In these cases, a patient may not have the time to become acclimated to a young psychotherapist when s/he prefers an old psychotherapist. In these situations, therapist characteristics which will help to facilitate a therapeutic relationship quickly are even more important (Landes et al. 2013). Finally, online platforms should use age group as a search criterion for psychotherapists.

Appendix

Development of study materials

Pilot 1

The aim of the first pilot study was to select pairs of naturalistic photographs of an old and a young psychotherapist. To ensure the validity of the photograph material, 120 randomly selected photographs of real psychotherapists published on the internet platform www.therapie.de were chosen (informed consent by the psychotherapists was obtained). The photographs were then rated by 31 participants (M = 27.34 years old, SD = 8.57) according to the chronological age of the psychotherapist. Psychotherapists rated as 35 years or younger (‘young psychotherapists’) as well psychotherapists rated as 55 years or older (‘old psychotherapists’) were selected for the study. 58 photographs fulfilled these criteria and were then rated on a five-point non-verbal pictorial assessment scale according to (1) valence and (2) arousal, using the self-assessment manikin (Bradley and Lang 1994). To identify matched young-old pairs of hypothetical psychotherapists, photographs with the smallest difference rating scores (Euclidean distance) on valence and arousal were selected. All Euclidean distances were small and ranged from d = 0.04 to 0.10.

Pilot 2

In the second pilot study, a pool of 15 fictitious scenarios representing various presenting problems were developed by the first and second author of this study, building on the research of Berlin Wisdom paradigm (Baltes and Staudinger 2000) and studies on patient preferences (Harris and Busby 1998). Each hypothetical presenting problem (three to four sentences long) was evaluated by a group of 25 participants (master students of a clinical psychology study program, M = 26.34 years old, SD = 4.81) on a five-point scale (1 = low, 5 = high) according to two dimensions, namely (1) to the degree to which the presenting problem described was developmentally close to and specific for young adulthood in today’s world as well as (2) to the degree to which the presenting problem was a universal problem addressing fundamental questions of life virulent across the adult life span. To be selected as a ‘young problem,’ a score larger than 3 on the first dimension as well as a significant difference between the two scores was required (difference t-values: 3.16, 2.41, 2.56, p < .05). The same criteria were applied for the ‘universal problem’ (difference t-values: − 2.70, − 4.31, − 3.61, p < .05). This procedure resulted in the three ‘young problems’ (cyberbullying, paternalizing parents, lovesickness) as well as the three universal problems (a friend’s death, natural disaster, quarreling with one’s past) as used in the main study. (In order to exclude confounding effects, each presenting problem was additionally rated by the same participants according to the degree of severity on a five-point scale (1 = minor, 5 = very severe). A problem was rated a priori as very severe in case of a score > 4, as severe-to-moderate in case of a score of 3 to 4, and minor-to-moderate in case of a score < 3. The degree of severity was equally distributed across presenting problems, i.e. both the category of ‘young problems’ and ‘universal problems’ each included a very severe problem (a friend’s death M = 3.2, cyberbullying M = 3.4), a severe-to-moderate problem (natural disaster M = 3.0, paternalizing parents M = 3.0), and a minor-to-moderate problem (quarreling with one’s past M = 1.8, lovesickness M = 1.5).)

Cyberbullying (‘young problem’)

When you take a shower after the sports program which is offered by the university, a fellow student takes pictures of you with his smartphone. When you come home and turn on your computer, you see that the student has shared the photograph on Facebook. Within the next few weeks, you notice that people start to talk about you behind your back during the lectures. Even after weeks, the photograph is still online.

Paternalizing parents (‘young problem’)

Although you have moved out from your parents a while ago, your parents enjoy staying in touch with you a lot. Increasingly, they become involved in your personal life. They interfere with your everyday life, your eating habits and vacation planning. In the meantime, you even feel guilty about not asking your parents for advice in certain situations.

Lovesickness (‘young problem’)

After signing up for the dating app Tinder, it takes you a while until you finally start to trust in the people you meet there. You get along very well with one of your ‘matches,’ and the two of you meet several times. You notice that you start to fall in love; however, you do not exchange your mobile numbers on any of your dates. After your last date, you discover that your heartthrob has unmatched you on tinder so that you can’t contact the other person anymore. You suffer from a broken heart.

Natural disaster (‘old problem’)

During your long-awaited trip to Indonesia, a severe earthquake happens. Nothing happens to you and your family. However, the village where you spent your holiday is badly damaged. Since then you are being tormented by thoughts about the families who lost their homes.

A friend’s death (‘old problem’)

One day, you return from your holiday. You get the news that a good friend of yours died in a car accident. Previously, you and the friend who died argued about something that now seems completely irrelevant to you.

Quarreling with one’s past (‘old problem’)

One evening, you feel lonely. After a feedback from a friend, you think back to a partner that you broke up with for no obvious reason. From then on, you begin to quarrel with your past and question the decisions you have made in your life.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

This study follows the principles of the Declaration of Helsinki.