Ruminative Responses Scale (RRS)

The Ruminative Responses Scale (RRS), originally developed by Nolen-Hoeksema and Morrow (1991), is a 22-item self-report measure of rumination in adults aged 18 and over. Rumination is a way of responding to low mood that involves repetitively and passively thinking about the causes, meanings, and consequences of distress rather than taking action to address it. The RRS asks respondents how frequently they engage in these ruminative responses when they feel down, sad, or depressed.

Ruminative Responses Scale Subscales

The RRS yields a total score and two subscales that distinguish a passive, self-critical form of rumination from a more deliberate, analytic one. The two subscales were derived by Treynor et al. (2003) to address a limitation of the original scale. Because several RRS items overlap in content with symptoms of depression, the authors set those items aside and examined the remaining ten, which were grouped into the Brooding and Reflection subscales.

The resulting 10-item Brooding and Reflection subscales have been widely adopted in subsequent research, whereas NovoPsych administers the original 22-item RRS and reports both the original total score and the Treynor-derived subscale scores. The remaining 12 items (not included in the subscales) assess rumination focused on depressive symptoms themselves and contribute only to the total score.

• Brooding (5 items): captures moody pondering, a passive dwelling on the gap between one’s current situation and an unachieved standard (for example, thinking “Why do I have problems other people don’t have?”).
• Reflection (5 items): captures a deliberate turning inward to analyse and understand low mood (for example, “Go away by yourself and think about why you feel this way”).

The Significance of Rumination

Rumination is one of the most well-established psychological processes that maintain and worsen depression. When people respond to low mood by repeatedly dwelling on how they feel, their symptoms, and the possible causes and consequences of their distress, rather than engaging in more adaptive coping strategies, that low mood tends to deepen and last longer. This happens because rumination amplifies the effect of negative mood on thinking, interferes with problem-solving, and can wear down social support. Prospective studies have shown that individuals who ruminate tend to experience more severe and longer-lasting depressive symptoms over time, supporting rumination as an important maintaining process in depression (Nolen-Hoeksema and Morrow, 1991; Nolen-Hoeksema et al., 1999). Rumination also helps explain why women experience depression more often than men, as women tend to ruminate more (Nolen-Hoeksema et al., 1999). Although it is studied most in depression, rumination is also implicated in anxiety and post-traumatic stress, which makes it a useful process to measure and track across a range of presentations.

Clinical Applications of the RRS

The RRS is among the most widely used measures of rumination in clinical research and practice. It is used in clinical settings wherever rumination is a suspected maintaining process, most prominently in depression. It is commonly used alongside rumination-focused cognitive behavioural therapy and metacognitive approaches, in which rumination is a primary treatment target.

In practice, the RRS can support several clinical tasks.

• When rumination is prominent, it can become a focus of therapy in its own right, and the RRS helps establish whether this is the case for a given client.
• The RRS supports case formulation whenever repetitive negative thinking appears to be holding a presentation in place.
• Sharing a client’s Brooding and Reflection profile helps convey that not all dwelling on problems is the same.
• The RRS can be administered repeatedly across therapy to see whether a client’s reported rumination is shifting, with change best read as a trend across several administrations rather than from any single comparison.

The measure’s most distinctive clinical contribution is the distinction between Brooding and Reflection. Brooding is the component most strongly associated with current and future depressive symptoms (Treynor et al., 2003). Reflection was originally found to accompany low mood in the short term while being linked to less depression over time, though later research suggests it is not reliably adaptive and may itself become unhelpful when it becomes repetitive or difficult to disengage from (Tang et al., 2021; Townshend and Hajhashemi, 2025). Considering the relative elevation of the two subscales can therefore help inform case formulation and guide treatment planning.

Ruminative Responses Scale Scoring

The RRS can be scored two ways. The original approach sums all 22 items into a single total score. A later approach (Treynor et al., 2003) scores only 10 of those items as two subscales. NovoPsych administers the full 22-item measure and reports both, so a completed RRS produces three scores, a 22-item Total Score plus the subscale scores. The remaining 12 items count toward the Total Score only.

The RRS Total Score is the sum of all items and ranges from 22 to 88, with higher scores indicating more frequent rumination. The 10-item scoring approach yields two subscales:

• Brooding (5 items; items 5, 10, 13, 15, 16; scores range from 5 to 20) captures moody pondering and passive negative comparison. Elevated scores identify the form of rumination most strongly associated with current and future depressive symptoms.
• Reflection (5 items; items 7, 11, 12, 20, 21; scores range from 5 to 20) captures a deliberate, analytic turning inward to understand low mood. Elevated scores indicate a more deliberate, analytic inward focus. This is less consistently tied to poor outcomes than Brooding, but should not be assumed protective.

Average Scores. The report also displays the average score for the total score and each subscale, calculated by dividing the score by its number of items (the total by 22, each subscale by 5). This puts every scale on a common 1 to 4 range that matches the four response options, and the frequency descriptor is applied to this average.

RRS Frequency Descriptors

The report applies a frequency descriptor to the RRS average score for the total score and each subscale to aid interpretation. The RRS has no established clinical cut-offs; these descriptors are not validated thresholds. NovoPsych created them from the scale’s response anchors, which ask how frequently the respondent thinks or does each ruminative response (almost never, sometimes, often, almost always), so each descriptor indicates how frequently the respondent ruminates on average, not how severe or clinically significant that rumination is.

The four frequency descriptors, and the average score ranges they cover, are set out below:

• 1.00 to 1.49: Almost never. The respondent reports, on average, rarely engaging in these ruminative responses when feeling down, sad, or depressed.
• 1.50 to 2.49: Sometimes. The respondent reports, on average, occasionally engaging in these ruminative responses when feeling down, sad, or depressed.
• 2.50 to 3.49: Often. The respondent reports, on average, frequently engaging in these ruminative responses when feeling down, sad, or depressed, suggesting rumination is a regular feature of their response to low mood.
• 3.50 to 4.00: Almost always. The respondent reports, on average, engaging in these ruminative responses almost all of the time when feeling down, sad, or depressed, suggesting rumination is a habitual response style.

Interpreting the RRS Subscale Profile

Looking at the balance between the Brooding and Reflection subscale scores, rather than either score alone, can indicate which form of rumination is more prominent for a client. When Brooding is the more elevated subscale, the client’s rumination is likely characterised more by passive negative comparison, the component most strongly associated with the persistence and worsening of depressive symptoms, and this may support considering rumination itself as a treatment target. When Reflection is the more elevated subscale score, the client’s inward focus has a more deliberate, problem-solving character. Reflection is less consistently tied to poor outcomes than Brooding, but an elevated score should not be assumed protective, as reflection is not reliably adaptive and has been linked to poorer outcomes in some clinical samples.

Reflection is best used as a pattern indicator rather than a severity score, and four points follow from that. First, the comparison between Brooding and Reflection carries the clinical signal, not either score by itself: a client high on Brooding with little Reflection is dwelling passively and self-critically, whereas a client high on both is working hard to analyse a low mood without that effort paying off, which points towards approaches targeting the process of rumination rather than its content. Second, a high Reflection score should not be treated as credit against a high Brooding score; in people who are currently depressed, reflection often functions much as brooding does, so it does not offset it. Third, Reflection describes how a client is likely to engage with rumination-focused work rather than how unwell they are: someone scoring highly is already turning the problem over analytically, so the task is usually to redirect that effort rather than to encourage more of it. Fourth, when tracking change across administrations, a fall in Brooding is the outcome to look for; a fall in Reflection is not in itself good news, and is best read alongside the Brooding trajectory.

Reviewing the most highly endorsed RRS items, particularly on the Brooding subscale, can help identify the specific ruminative content a client returns to, which can inform therapeutic treatment.

RRS Graphs and Change Over Time

The report is arranged in two blocks. The first carries the total score: its results table, its own graph, and a note on what the total contains. The subscales then begin on a new page with their own results table and a graph displaying Brooding and Reflection together, so the two can be read against each other. The total is kept separate because it is based on all 22 items, including the 12 that are not part of either subscale.

Every graph plots the average score and labels all four frequency descriptor bands, which are shaded as a graded scale from the lowest band (unshaded) to the highest. On the Total and Brooding graphs the bands are shaded so that stronger shading marks a higher frequency of rumination and greater clinical concern; on Reflection those bands take a neutral shading that marks how frequently the respondent reflects without implying that a higher score is necessarily of more clinical concern.

On a single administration, the RRS scores are shown as bar charts, and across repeated administrations they are displayed as line charts plotting change over time. Change in RRS scores is described in plain descriptive terms rather than as a statistically meaningful change score, because the RRS has no established threshold for reliable or clinically meaningful change. Rumination scores in untreated community samples reduce modestly on re-administration (Treynor et al., 2003), so small decreases are best read as part of a trajectory across several administrations rather than as treatment effects in themselves.

Development of the Ruminative Responses Scale

The RRS began as the 22-item ruminative responses subscale of the Response Styles Questionnaire (RSQ), developed by Nolen-Hoeksema and Morrow (1991) to measure the ruminative response style described in Response Styles Theory (Nolen-Hoeksema, 1991). The theory proposes that people who respond to sad or depressed mood by ruminating experience more severe and longer-lasting depressive symptoms than people who respond by distracting themselves or taking action.

The RRS subscale structure was established by Treynor et al. (2003) in a secondary analysis of data from a randomly selected community sample of adults in the United States interviewed twice over one year (N = 1,328 at the first interview; 1,130 provided data at both waves; Nolen-Hoeksema et al., 1999).

To address concerns that some RRS items overlap in content with depressive symptoms, the authors excluded 12 symptom-related items and subjected the remaining 10 to principal components analysis. This revealed two five-item components, Brooding and Reflection, together accounting for 50.5 percent of the variance (Treynor et al., 2003). It is worth being explicit that the total score and the subscales therefore rest on slightly different measurement logics: the 22-item total is the classic, content-broad index and correlates with depression partly through shared item content, whereas Brooding and Reflection were constructed specifically to strip that overlap out. In this sample, the average Brooding score was 9.40 (SD = 2.96) and the average Reflection score was 9.83 (SD = 3.11) at the first interview. Although the present profile describes the original 22-item RRS, the Brooding and Reflection subscales derived from this analysis have become the most commonly used scoring approach in subsequent research.

RRS Reliability

Internal consistency in the community sample first reported by Nolen-Hoeksema et al. (1999) and reanalysed by Treynor et al. (2003) (N = 1,130) was strong for the 22-item RRS total score (α = .90) and acceptable for the five-item subscales (Brooding α = .77; Reflection α = .72). Across a one-year interval, test-retest correlations were .67 for the total score, .62 for Brooding, and .60 for Reflection, consistent with rumination behaving as a relatively stable response style that nonetheless shifts over time.

RRS Factor Structure

The RRS’s two-factor Brooding and Reflection structure is well supported. In the largest test to date, a confirmatory factor analysis of pooled data from 4,205 university students supported the two-factor model over a single-factor alternative, with one error covariance between two similarly worded Reflection items (robust CFI = .951, RMSEA = .065; Whisman et al., 2020). The same study demonstrated configural, metric, and scalar measurement invariance across gender, indicating that Brooding and Reflection scores carry the same meaning for women and men and can be meaningfully compared.

At the level of all 22 items, exploratory analyses show a broad general rumination factor with a distinct Reflection component (Treynor et al., 2003); the Brooding and Reflection subscales are defined within the refined 10-item subset.

RRS Validity

The 22-item RRS total score correlated .48 with concurrent depressive symptoms on the short-form Beck Depression Inventory (BDI) and .38 with depressive symptoms one year later (Treynor et al., 2003). The two subscales (Brooding and Reflection) relate to depression differently, and this differential pattern is the core of the measure’s validity evidence. Brooding correlated .44 with concurrent depressive symptoms and predicted greater depression one year later after controlling for baseline symptoms; Reflection correlated .12 concurrently, and its zero-order correlation with depression one year later was also positive (.08); a negative coefficient emerged only in a multivariate model, which the authors noted may reflect statistical suppression (Treynor et al., 2003). Brooding was also more strongly associated than Reflection with a lower sense of mastery (r = -.26) and higher chronic strain (r = .20).

As predicted by Response Styles Theory, women scored modestly higher than men on both subscales (d = .16 to .22; Treynor et al., 2003), and Brooding, but not Reflection, statistically mediated the gender difference in depressive symptoms. Latent-mean comparisons under scalar invariance confirmed small but significant gender differences in the same direction (Whisman et al., 2020).

RRS Norms

Community data for the RRS-22 does exist (Nolen-Hoeksema et al., 1999). That sample was administered the RRS via face-to-face interview rather than self-report, was recruited in the 1990s from three Californian cities in three narrow age bands that exclude adults aged 18–24, 36–44 and 56–64, and produced no scores above 76 of a possible 88. Percentiles derived from it would imply a reference comparison that does not hold for the people completing the RRS via self-report in NovoPsych, particularly at the high end of the scale. As such, NovoPsych does not calculate RRS percentiles.

Assess your patients with the Ruminative Responses Scale (RRS) using a NovoPsych account

• For Psychologists & Mental Health Clinicians
• Send Assessments to Patient’s Phone
• Receive Comprehensive Reports
• Access Over 100 Validated Assessments​
• Instant Psychometric Scoring
• Track Symptoms​
• Inform Treatment​