This study used cross-sectional data gathered for the Mobility and Active Aging study (MIIA) conducted at the University of Jyväskylä, Finland. The present participants were randomly selected from among the 848 participants of the Life-Space Mobility in Old Age (LISPE), which was a larger population-based study with a probability sample from the national population register . The present sample was planned to comprise only part of the original LISPE sample, since the power calculations showed that a sample of 200 persons would be sufficient for statistically significant, moderate correlations. Thus, 298 persons were invited to participate. Of these, 77 declined to take part and 15 were not reached. Thus, the present data were gathered from 206 community-dwelling older adults who were aged 79–93 years, able to communicate and living independently in the Central Finland municipalities of Jyväskylä and Muurame.
Of the 848 LISPE participants, the present 206, who also participated in the MIIA study, were somewhat younger (80.0, standard deviation SD 4.1, vs. 80.8, SD 4.3, p = 0.02), and had slightly better cognition (Mini Mental State Examination, 26.6, SD 2.3, vs. 26.0, SD 2.9, p = 0.01) and physical performance (Short Physical Performance Battery, 10.2, SD 1.8, vs. 9.5, SD 2.7, p < 0.001) than the others (n = 642). The LISPE + MIIA and LISPE-only participants did not differ by sex, number of chronic conditions, or years of education.
The present data were collected by computer-assisted face-to-face home interviews in spring 2016. In total, 186 participants answered the questions on TGP, FGA, perceived autonomy and life-space mobility. The study protocol was approved by The Ethical Committee of the University of Jyväskylä. Participants signed informed consents before the assessments.
Tenacious goal pursuit and flexible goal adjustment
Coping was assessed with short versions of the Tenacious Goal Pursuit (TGP) and Flexible Goal Adjustment (FGA) scales, originally developed by Brandtstädter and Renner . The short versions of the scales each contain five items, such as ‘Even when things seem hopeless, I keep on fighting to reach my goal’ (TGP) and ‘If I do not get something I want, I take it with patience’ (FGA) . The response options are consistent with a five-point Likert scale: ‘strongly agree’ (0), somewhat agree (1), doesn’t agree or disagree (2), somewhat disagree (3), and ‘strongly disagree’ (4). There is one inversely phrased item in both scales. Henselmans et al.  reported weak face validity for the inversely phrased items. In the present study, the Cronbach’s alphas were higher when the inversely phrased items were omitted (TGP: α = 0.77 without vs. α = 0.72 with the inversely phrased item, FGA: α = 0.67 vs. α = 0.60). Moreover, the correlations between the directly and inversely phrased scores were rather low (TGP: r = 0.02–0.18, FGA: r = 0.09–0.21). Thus, we omitted the inversely phrased items. The remaining four items in both scales were reverse-scored with higher scores indicating higher tenacity or flexibility, and a sum score (range 0–16) was calculated for each scale when responses were given to at least three of the four items. Single missing items were imputed with the mean of the existing values of the respective participant (n = 3 in TGP, n = 3 in FGA). We excluded 20 participants from the analyses, since they had not responded to any of the questions concerning coping. These participants had lower cognition than the participants included in the analysis (MMSE mean 23.9 vs. 26.1, respectively) and had missing data also in the depression questionnaire (CES-D). We could not use any other time points to estimate responses for these participants.
Perceived autonomy in participation outdoors
The ‘autonomy outdoors’ subscale of The Impact on Participation and Autonomy (IPA) questionnaire was used to assess perceived autonomy in out-of-home activities. The IPA is a validated measure, which can be used as a whole or in part (subscales) to assess participation and autonomy [26, 27]. The ‘Autonomy outdoors’ subscale comprises five items on perceived possibilities to (1) visit relatives and friends, (2) make trips and travel, (3) spend leisure time, (4) meet other people, and (5) live life as one wants. The response options range from ‘very good’ (0) to ‘very poor’ (4). A sum score (range 0–20) was calculated with higher scores indicating poorer autonomy.
Life-space mobility refers to the spatial area an individual purposely moves through in daily life. It factors in all movement irrespective of the mode of transportation and reflects person’s access to community amenities. It was measured with the Finnish version  of the University of Alabama at Birmingham Study of Aging Life-Space Assessment (LSA) . The assessment includes six life-space areas starting from the informant’s bedroom and expanding to include the home, yard, neighborhood, town, and beyond town. Participants are asked how often they have moved in each area during the 4 weeks preceding the assessment and whether in doing so, they have needed help from any devices or another person. In the analyses, we used a life-space mobility composite score, which reflects the distance, frequency, and level of independence of mobility, with higher scores (range 0–120) indicating higher life-space mobility . The reliability and validity of the LSA measurement have been established [6, 10].
In addition to age and sex, which were drawn from the national population register, objectively measured physical and cognitive performance and entrance-related environmental barriers were regarded as theory based confounders. Cognitive performance was assessed with the Mini Mental State Examination (MMSE)  and lower extremity function with the Short Physical Performance Battery (SPPB) . Environmental barriers at entrances and in close exterior surroundings were objectively recorded using the Housing Enabler screening tool [30, 31]. Depressive symptoms were assessed with the Centre for Epidemiologic Studies Depression Scale (CES-D) .
Morbidity was evaluated as the number of self-reported physician-diagnosed chronic conditions from a list of 22 diseases including, e.g. coronary artery disease, diabetes, cancer, and Alzheimer’s disease. An additional open question was asked about conditions other than those on the list . Years of education was also self-reported.
First, the correlations between TGP, FGA, life-space mobility, and perceived autonomy in participation outdoors were tested with Pearson’s correlation. Thereafter, to identify the different coping profiles in our sample, we performed a cluster analysis of TGP and FGA using two-step clustering. Two-step clustering identifies groupings by first running pre-clustering and then running hierarchical methods. Log-likelihood was used as a distance measure and the number of clusters was not determined beforehand. Since cluster solutions can depend on the order of cases, the order was randomized before the analysis. To test the stability of the given solution, cluster analysis was executed four additional times using different randomizations of cases. Differences in background characteristics between the resulting coping profiles were analyzed with chi square test and one-way analysis of variance.
Finally, general linear modeling was used to study the associations of the coping profiles with life-space mobility and perceived autonomy in participation outdoors. The base model was adjusted for age and sex. SPPB, MMSE, environmental barriers, and depressive symptoms were added to the base model one at a time to see which one of them possibly affects the associations. All analyses were performed with SPSS Statistics 24 for Windows.