We Are Family: Fathers' Time with Children and the Risk of Parental Relationship Dissolution

Introduction

Over the past forty years, the role of fathers in children's lives, and the meaning of fatherhood itself, has changed substantially (Coltrane 1996, 2004; Doucet 1996; Gerson 1993; Miller 2010; Pleck 2010; Pleck and Pleck 1997). Rising rates of maternal employment and greater appreciation of fathers' role in child development have popularized a more time-intensive fatherhood ideal (Daly 2001), one that prescribes more “nurturing” parenting roles than the traditional father-as-breadwinner or -disciplinarian (Coltrane 1996, 2004; Gerson 1993; Pleck and Masciadrelli 2003). Between 1965 and 1985, fathers in the United States averaged about 2.5 hours per week in active childcare, but by 2000 this figure had nearly tripled to almost 7 hours per week (Wang and Bianchi 2009). At the same time, a rise of intensive parenting norms has driven increases in mothers' active time with children (Dew 2009: Ramey and Ramey 2010), and it is still the case that fathers spend significantly more time in paid work than mothers and do less parenting and housework than their wives (Coltrane 2004).

These shifting patterns of time use are likely to affect marital and family relationships. Mounting evidence shows that the quality and quantity of father involvement is positively related to children's outcomes (Pleck 2010). The present paper examines how fathers' time with children is associated with parental relationship dissolution. Despite intense interest in fathers' time with children, we know of no published studies that analyze how such time correlates with relationship dissolution, though a number of studies have examined fathers' (or men's) participation in housework as a correlate of marital dissolution and marital satisfaction (e.g., Coltrane 2000; Cooke 2006; Hochschild and Machung 1989; Weagley, Chan, and Yan 2007).

This study aims to understand whether variation in the amount, composition, and type of fathers' time with children is associated with the risk of relationship dissolution. We examine these questions using data from the Longitudinal Survey of Australian Children (LSAC), which is a representative longitudinal panel survey of approximately 10,000 Australian families with young children. The LSAC includes time-diary data from the child's perspective as well as rich survey measures of the parental relationship and family background characteristics. An important feature of these data is that they allow us to identify the composition of fathers' time; specifically, by distinguishing fathers' time alone with children from time with children and the mother together. In this paper, we refer to the former as “solo time” and the latter as “shared time.” Moreover, we are able to measure a broad variety of kinds of fathers' time with children, including eating, watching television, basic care, management, teaching, and play. Finally, the data allow us to construct measures of parental affection and maternal well-being. Doing so allows us to explore potential mechanisms linking fathers' time with children and the likelihood of relationship dissolution.

Background

Perspectives on Fathers' Time with Children and the Risk of Relationship Dissolution

Fathers' time in childrearing can be characterized in terms of its amount, composition (i.e., shared versus solo), and type (i.e., routine activities such as basic care and management versus potentially more fun activities such as play). We describe each of these dimensions of time and formulate ideas about how time characterized in these different ways is likely to correlate with relationship dissolution.

First, the amount of time that fathers contribute to childrearing is expected to correlate with the risk of relationship dissolution. One possibility is that fathers' higher levels of time investment in childrearing will improve maternal well-being, thereby lessening the risk of relationship dissolution. This could be because fathers' spending time with children demonstrates their emotional engagement in the marriage and affection for their spouse (especially if the time with children is also spent together with the spouse). According to Wilcox and Nock (2006), a husband's “marital emotion work,” that is, his emotional engagement, love, affection, and understanding, is the primary predictor of women's reported marital quality. Fathers' time with children may also lessen mothers' time pressure, which, by reducing her parenting stress and depression (Folbre et al. 2005), strengthens the relationship. Moreover, because contemporary dual-earner marriages require both spouses to invest time at home, fathers' time with children may signal commitment to the family and this might lead to a happier marriage (Wilcox and Nock 2006). In sum, this suggests that the higher the level of fathers' time investment, the lower the risk of relationship dissolution.

Second, the composition of fathers' childrearing time may correlate with relationship dissolution. Prior studies have shown that it is important to distinguish fathers' and mothers' time alone with children versus time with children that the mother and father spend together (Folbre et al. 2005). In fact, mothers are also present during most of the time that fathers spend with their children (Craig 2006; Craig and Mullan 2011; Fuligni and Brooks-Gunn 2000; Kalil, Guryan, and Ugaz 2010). Craig and Mullan (2011) also showed that higher levels of father education correlate with a greater share of father solo parenting time in Australian time-use data.

It is unclear whether fathers' solo or shared parenting will be more relevant for relationship dissolution. On one hand, mothers report that “meaningful” sharing of childrearing is characterized only by fathers' solo time investment (as opposed to shared time; see Fuligni and Brooks-Gunn 2000). This underscores the idea of reduced maternal time pressure or parenting workload as an important outcome of increasing levels of fathers' time investments. If reduced maternal time pressure is an important correlate of relationship quality and dissolution, then fathers' solo time may be particularly important.

On the other hand, a large literature suggests that shared parenting time may be a key venue for emotional engagement and as such may be an important correlate of relationship dissolution. Family systems theory suggests that shared parenting time can serve as a forum for building strong relationships that help promote the marriage (Cox and Paley 1997). For instance, shared time might allow parents to support and learn from each other's parenting strategies, and this could build marital stability (Van Egeren and Hawkins 2004).

Discussions of shared time have also entered the economics literature. Hamermesh (2002), for example, argued that couples prefer to consume leisure simultaneously. Similarly, Stevenson and Wolfers (2008) argued that the key to successful marriage in the contemporary era is “consumption complementarities—activities that are not only enjoyable but are more enjoyable when shared with a spouse” (2). Data from the American Time Use Survey show that couples' joint time in leisure activities has increased over the past four decades, which suggests that consumption complementarities are increasingly important in US marriages (Tabasso 2011).

Furthermore, Lesnard (2008) described shared parenting time with children, specifically in eating and leisure, as “the core of contemporary family solidarity,” arguing that shared family time, especially in activities that promote relaxation and amusement, not only represents the fabric of everyday family life but is also arguably the main forum for couples to express their love and affection for each other. Others have similarly argued that shared family mealtimes provide key opportunities for communication, and that it is this aspect of the family meal in particular that is associated with family members' positive affect, well-being, and relationship building (Fulkerson, Neumark-Sztainer, and Story 2006; Ochs and Shohet 2006; Offer 2013a, 2013b). In sum, there are good reasons to think that both solo and shared time are important correlates of relationship dissolution.

Third, the type of activities fathers engage in with their children may correlate with relationship dissolution. Numerous studies have shown that fathers engage in different types of parenting activities than mothers. Fathers spend a larger proportion of their parenting time in “fun” play and companionate activities like sports, games, roughhousing, and pretending, whereas mothers spend a larger proportion of their time in “routine” caregiving and nurturing, particularly with young children (Craig 2006; Craig and Mullan 2011; Pleck and Pleck 1997). The fact that play and companionship characterize fathers' child time in greater proportion to mothers' time suggests that fathers serve a distinct parenting role. However, Fuligni and Brooks-Gunn (2000) noted that mothers regard only father participation in routine (non-leisure) care activities as constituting meaningful sharing of parenting. Mothers seem to hold the view that fathers' “recreational” activities with children are less work than they are leisure. This suggests that fathers' participation in routine parenting tasks, such as basic care and management, plays a stronger role in relationship dissolution than activities such as father playtime. At the same time, as noted above, fathers' time in family meals is also expected to be a key activity.

Methodological Challenges

Key concerns affecting our investigation include selection bias and reverse causality. One can imagine numerous factors that predict both the likelihood of fathers' involvement in childrearing time and the risk of relationship dissolution. For instance, parental education and employment may affect the amount, composition, or type of parenting time (Craig and Mullan 2011) in addition to relationship dissolution. Moreover, a variety of scholars discuss men's views of work, family, and marriage as a “package deal,” with father-child engagement shaped substantially by their relationships with their wives or partners as well as by their work lives (Coltrane 2004; Townsend 2002). Also consistent with reverse causality, Schmierer (2010) uses data from the Children of the NLSY-79 to document that fathers' time with children decreases in the years leading up to divorce, a pattern he attributes to fathers' anticipation of the divorce. We address these concerns by investigating robustness to controlling for a wide range of observable demographic characteristics as well as baseline measures of marital quality and happiness.

Data and Method

Institutional Setting

The setting for our study is Australia 2004–2008, a country in which men's and women's attitudes about sex roles in the household division of paid and unpaid work are similar to those in the United States. In neither country do men or women have particularly egalitarian attitudes toward the distribution of household and labor-market work (van Egmond et al. 2010). Van Egmond et al. (2010) report that from 1986 to 2005, egalitarian views in Australia slowed or even reversed. The rate of maternal employment among those with children below the age of 15 is nearly identical in the United States and Australia, at approximately 67 percent in 2008 (OECD 2011), though Australian mothers are less likely than US mothers to work full-time when their children are young. Australia has a high share of households in which fathers work full-time and mothers work part-time (Craig and Mullan 2011).

Trends in fathers' time involvement in Australia have largely mirrored those in the United States, with large increases documented in paternal childcare time since the mid-1980s (Bianchi, Robinson, and Milkie 2006). Intensive, hands-on parenting among more educated parents is normative in Australia, as it is in the United States (Craig and Mullan 2011). Though Australia's family wage policy is generally indicative of support for a male breadwinner model, it was also the only country to have comparable worth policies until the mid-1990s. At that time, the Australian gender wage gap of 12 percent was among the smallest in industrialized societies. By way of comparison, the gap in the United States was 22 percent at that time (Harkness and Waldfogel 1999). Finally, in both the United States and Australia, divorce rates increased over recent decades to a peak in 2008 of 3.7 and 2.3 divorces per 1,000 people, respectively (OECD 2011).

Data

The Australian Institute of Family Studies' Growing Up in Australia, the Longitudinal Study of Australian Children (LSAC), follows two birth cohorts—the B cohort of 5,107 children born between March 2003 and February 2004, and the K cohort of 4,983 children born between March 1999 and February 2000. In the first wave of data collection in 2004, children in the B cohort were between the ages of zero and one, and children in the K cohort were between the ages of four and five. The study has interviewed families biennially since 2004, and has mailed questionnaires to families annually in between-wave years. The Australian Institute of Family Studies has released waves 1–4 of data collection as of March 2012.

The study aims to understand how a child's social, economic, and cultural environments shape development and well-being. The LSAC collects information about children's cognitive skills, socioemotional functioning, physical health, education, and family and community contexts through face-to-face interviews, self-administered questionnaires, time diaries, and assessments (AIFS 2011). There are many benefits to using these data for the present study, including a large sample that is homogeneous with respect to child age, rich longitudinal survey data, time-diary data gathered from the child's perspective (which allows for a distinction between parents' shared and solo time), and an objective marker of family life (e.g., relationship dissolution).

Analytic Sample

We use a pooled sample containing children from both of the LSAC's cohorts, following the children across two periods of data collection (periods 1 and 2). To construct a pooled sample of comparably aged children, period 1 (or baseline) consists of data from the B cohort's wave 2 survey and from the K cohort's wave 1 survey, and period 2 consists of data from the B cohort's wave 3 survey and from the K cohort's wave 2 survey. Our study therefore follows a group of children starting from when they are aged two to five in period 1 and ending when they are aged four to seven in period 2. From the original sample of approximately 10,000 children, 90 percent of the K cohort and 86 percent of the B cohort remained in the study at period 2.

Period 1 time-diary response rates were 78 percent for the K cohort and 75 percent for the B cohort.¹ Among children with available and useable² time-diary data, we exclude those who have zero minutes of total parenting time in period 1. These children comprise 5 percent of the K cohort and 2 percent of the B cohort. In order to investigate the association between parents' time in childrearing in period 1 and parental relationship dissolution by period 2, we necessarily restrict the pooled sample to children who live with two parents. We restrict this sample to children with two opposite-sex biological or adoptive parents who live together in the child's home, report themselves as being partners at period 1, and provide baseline interview and survey data from which we can derive child and family background measures. Among the remaining 5,772 children, we dealt with outliers by trimming the top percentile on total solo mother care time (1,108 minutes), total solo father care time (557 minutes), or total shared time (1,258 minutes) in period 1. All told, our restriction process leaves us with an analytic sample of 5,604 children.

In addition to our main analyses predicting relationship dissolution as a function of parents' time in childrearing, we perform additional analyses that use alternative outcome measures (period 2 measures of parental relationship quality and happiness, maternal stress and depression) in order to investigate the mechanisms by which parents' time in childrearing might relate to relationship dissolution. The sample size for these analyses (shown in table 6) is smaller (4,650) because only children from the pooled sample with non-missing data on these outcomes are included.

Time-Diary Data

At each wave of data collection, the LSAC attempts to gather two time diaries from each study child, one referring to a weekday and another referring to a weekend day. An adult completes the time diary on behalf of the study child; in almost all cases, this is the mother. Adult respondents are left with several self-complete forms after their personal interviews, including two time diaries, and the interviewer works through an example diary before leaving. The interviewer tells the adult respondent which dates have been selected as the designated diary days, and tells them to complete the diary two or three times during the diary day, all at once when the child goes to bed, or in the morning after the diary day.

The diary day is divided into 96 fifteen-minute increments. The adult respondents are told to report what the “child has been doing over the 24 hours that started at 4 a.m. today and ends at 4 a.m. tomorrow,” including contextual information. For each time slot, the diary collects information about what the study child is doing, including whether and by what means the study child is traveling, where the study child is located, who else is present (“who was in the same room, or who was near the child if the child was outside”), and whether someone paid for the activity.

The LSAC time diary lists activities in which the study child might be engaged, and the adult respondent indicates in which of these activities the child participates for each of the 96 time slots. The respondent can indicate up to four activities. The diary does not collect information on which among multiple activities in a given time slot is the primary one, nor whether the child is engaging in multiple activities simultaneously or in succession. Thus, to avoid double-counting time in cases where multiple activities per time increment are recorded, we divide the 15-minute increment by the number of activities recorded. The only exception to this decision rule involves the definitions of eating time and television time. If both eating and television time are reported in a 15-minute increment, then this is recorded as a single activity referred to as “eating-while-watching-television.” We do this because it allows us to distinguish between interactive and non-interactive mealtime, which the previous literature suggests is important.

For parsimony, we have separated the pre-coded LSAC activities into the following categories: eating, watching television, basic care, management, teaching, play, other, and missing. These activities reflect key dimensions of children's time, as described in Yeung et al. 2001). More details on our activity categorization are presented in the appendix.

The LSAC time diary also lists people with whom the study child might be (alone; mother or stepmother; father or stepfather; grandparent(s) or other adult relative; brother(s), sister(s), or other children; other adult(s); dog, cat, or other pet (not fish)). The adult respondent indicates which of these people is in the same room as the child (or nearby for activities taking place outside) for each of the 96 time slots. The respondent can indicate multiple people. Note that our parenting time measures do not capture parental care if the parent is not in the same room as the child. Furthermore, the measures do not capture the intensity or quality of care when parents are with the child.

Time-diary data offer several advantages over stylized survey questions of parents' behavior. First, time diaries explicitly tie data to “clock” time and so are more precise with respect to the quantity of time. Second, research has shown that time-diary methods produce better estimates of time in non-work activities than stylized questions (Robinson and Goodbey 1999). Third, time-diary methods reduce social desirability bias because it is more difficult to fabricate any given activity situated in a chronological report of other activities (Dew 2009). Finally, time-diary methods are explicitly designed to identify who the participant was with during each activity. This makes the data uniquely suited to identifying parents' shared time with children.

Measures

Outcome measures

Our main outcome measure of interest is relationship dissolution. We consider the parental relationship to have dissolved if (1) either of the child's parents moves out of the home between periods 1 and 2; or (2) either parent reports having been divorced between periods 1 and 2; (3) either the mother or father claims that s/he no longer lives with the study child at period 2 because of a relationship breakdown (including trial separation); or (4) the parents are no longer partners at period 2 (but a death has not occurred).

We also consider four alternative outcomes that enable us to empirically investigate potential mechanisms by which parents' time in childrearing might lead to relationship dissolution. First, we measure relationship quality and relationship happiness at period 2. We measure relationship quality with an adaptation of the Hendrick Relationship Quality Scale (Hendrick 1988) as reported by the study child's mother. The original scale contains seven items, but the LSAC provides a six-item, five-point scale (shown in the appendix) that asks questions regarding relationship satisfaction. In order to measure general relationship happiness, we utilize a period 2 survey item that asks the mother, on a seven-point scale: Which best describes the degree of happiness, all things considered, in your relationship? We code both indicators such that higher scores indicate higher levels of relationship quality and happiness.

Second, we measure maternal stress and depression at period 2. Maternal stress is measured on a five-point scale by taking the mean across two survey items that ask the mother (1) How difficult do you feel your life is at present? and (2) How well do you think you are coping? Maternal depression is measured with the six-point, six-item K-6 Depression Scale (Kessler et al. 2002). Again, we code both indicators such that higher scores indicate higher levels of maternal stress and depression.

Parenting time measures

As described above, parents were asked to fill out two 24-hour time diaries tracking the study child's time use, one on a weekday and the other on a weekend day, at each wave of data collection. Following Bittman et al. (2003), we use an averaging process across the two diary days to create a single representative day, by multiplying the amount of time dedicated to each weekday activity by 5, multiplying the amount of time dedicated to each weekend-day activity by 2, adding the two, and dividing by 7. When children have two weekday diaries or two weekend-day diaries, we take a simple average across the diaries to create a single representative day, and when children have only a single diary, we use this day as the representative day.

Based on data from this representative day, we create three basic measures of parenting time in period 1: the total time the child spends alone with the mother (solo mother care), the total time the child spends alone with the father (solo father care), and the total time the child spends with both parents together (shared time). We then divide the total time in each of solo mother care, solo father care, and shared according to the activities undertaken. Specifically, we create measures of time spent in basic care, management, teaching, playing, television (without eating), eating without television, eating with television, “other,” and “missing.” “Missing” is for periods in which information is not provided regarding the activity undertaken. In some analyses we combine the two eating variables, and in others we separate them.

Child and family controls

All our models include controls for characteristics of the study child and his or her parents at baseline (period 1). We control for the study child's gender, age, birthweight, his/her number of siblings living in the home, whether s/he speaks a language other than English at home, and a measure of his/her temperament. We control for the mother's and father's ages, their immigrant and indigenous statuses, whether they have children living elsewhere, whether they regularly attend religious services, their levels of education, their marital status, and the length of their relationship. We also control for whether the mother filled in the time diaries. Some models control for characteristics of the parents' employment and earnings. More detailed information about these measures is available in the appendix. For our robustness checks (tables 4 and 5) and mechanism investigation (table 6), we also include period 1 measures of relationship quality, relationship happiness, maternal stress, and maternal depression (used as alternative outcomes when measured at period 2).³

Empirical Model

The parameters of interest in equation (1) are α_m, α_f, and α_s, which capture the impact of mother solo time, father solo time, and shared time on the likelihood of relationship dissolution. Estimates of equation (1) will produce unbiased estimates of these parameters, provided that there are no omitted determinants of the likelihood of relationship dissolution that are correlated with parenting time.

This identifying assumption is difficult to defend for several reasons. For example, parents who love each other are presumably more inclined to spend time together parenting and less likely to dissolve their relationship. Alternatively, depression may cause an individual to spend less time both parenting and investing in the partner relationship. In our empirical analysis, we address these potential sources of bias by investigating the robustness of our estimate to inclusion and exclusion of our large set of covariates capturing important parental characteristics, including mother's perception of relationship quality and maternal well-being. Admittedly, in the absence of experimental data, the threat of selection bias cannot be fully resolved. In particular, a reverse causality story is possible: higher expected probability of relationship dissolution may induce parents to decrease investments in children (Schmierer 2010).

Empirical Results

Summary Statistics

In table 1, we present summary statistics of our key outcome variable, relationship dissolution by period 2, in addition to the child, father, and mother characteristics in period 1, which we employ as controls in our regression analyses. We can see that 4 percent of the relationships observed in period 1 dissolve by period 2. With respect to education, 31 percent of the fathers and 35 percent of the mothers have completed a bachelor's degree, whereas 47 percent of the fathers and 43 percent of the mothers have not completed secondary education. Only four percent of the fathers are not employed, whereas 41 percent of the mothers are not employed. Almost 87 percent of the parents are legally married at period 1. Mothers' mean age is 35, and fathers' mean age is 37. The mean child age is 3.2 years, and on average a child has 1.35 siblings living in the home.

Table 2 presents summary statistics of our measures of parenting time. We can see that in a 24-hour period, a child spends on average 335 minutes in shared time, 325 minutes in mother solo care, and 69 minutes in father solo care.

Main Results

Table 3 presents our main findings. First, model 1 presents the results of a logit model regressing the probability of relationship dissolution by period 2 on total shared time, total mother solo time, and total father solo time, controlling for the rich set of background characteristics presented in table 1. The last column reports the marginal effect. There is a significant negative association between total shared time and the likelihood of relationship dissolution. The estimate is small, however, suggesting that a one-hour-per-day increase in total shared time will decrease the likelihood of relationship dissolution by .1 percentage points. Considering a dissolution rate of 4 percent, this suggests that a 30-minute-per-day increase in shared time will lead to a 1.25 percent decrease in the likelihood of relationship dissolution.

In model 2, we run the same logit model as in model 1; however, we employ the more detailed time-use categories. Interestingly, we can see that it is only shared eating time that is significantly associated with the likelihood of relationship dissolution. Moreover, the marginal effect estimate is large; a one-hour increase in shared eating time will decrease the likelihood of relationship dissolution by 2.3 percentage points. This suggests that a 30-minute-per-day increase in shared eating time leads to a 29 percent decrease in the likelihood of relationship dissolution.

As discussed in the empirical strategy, we should be concerned that the negative association between shared eating time and the likelihood of relationship dissolution estimated in table 3 is biased by unobservables. In particular, our model may have omitted determinants of the likelihood of relationship dissolution that are correlated with parenting time. In table 4, we carefully investigate how our estimate is robust to the stepwise inclusion of observable characteristics. First, in model 1, we run the same logit as in model 2 of table 3; however, we omit the child and parent covariates. We can see that the marginal effect estimate of shared eating time on likelihood of relationship dissolution is –2.9 percentage points. When we add child covariates in model 2, this estimate slightly increases to –3.0 percentage points. When we add controls for parental characteristics in model 3, then the marginal effect estimate slightly decreases to –2.4 percentage points. Finally, in model 4 we add controls for parent employment characteristics (detailed in table 1). This model should be interpreted with caution, as parent employment is endogenous to parenting time. Nevertheless, it is worth noting that the inclusion of these controls has no effect on our coefficient of interest. Comparing models 1–4, we can see that the estimated association between shared eating time and relationship dissolution is robust to the inclusion of a rich set of child and parent characteristics. Because parental labor-force participation is endogenous to parenting time, we will in the following use model 3 as our preferred model (identical to model 2 in table 3).

Also discussed in the empirical strategy, a particular concern for a causal interpretation of our estimate is that couples with a poor relationship are less inclined to spend time eating together as a family and presumably have a higher likelihood of relationship dissolution. Moreover, if the mother is stressed or depressed, the father may be less eager to spend time with the family. Some researchers posit that maternal stress and depression increase the likelihood of relationship dissolution (Davila et al., 1997). In table 5, we investigate these possible sources of bias by including controls for different measures of relationship quality and maternal well-being in period 1. First, for the sake of comparison, model 1 presents our preferred model. Model 2 is identical to model 1 but adds our two measures of relationship quality in period 1. We can see that both of these measures are strongly correlated with relationship dissolution by period 2. Nevertheless, including these controls does not affect the association between shared eating time and dissolution. In model 3, we add our measures of maternal stress and depression in period 1, and we can see that only maternal stress is predictive of relationship dissolution. But again, including these measures as controls barely affects our estimate of interest. In model 4, we also demonstrate that our estimate is robust to the inclusion of all four measures. Interestingly, once we control for relationship quality, maternal stress is no longer significantly predictive of relationship dissolution.

Mechanism Investigation

In models 1–4 of table 6, we further explore potential mechanisms by which parents' time in childrearing might lead to relationship dissolution. First, we can observe in our data that 6.5 percent of shared eating takes place during television watching (see table 2). Other studies have also suggested that a major distraction during family meals is television because it inhibits social interactions (Fiese and Schwartz 2008). Because sharing meals while watching television allows for fewer opportunities for emotionally engaged interaction compared to sharing meals without TV, we should see differential effects of shared eating time with and without TV. This would be consistent with Offer's (2013a, 2013b) argument that the key merit of the family meal vis–à–vis emotional well-being is the opportunity for family communication it provides. We address this in model 1 of table 6 by adding this distinction to model 4 of table 5. We find that it is only shared eating time without television that is negatively associated with relationship dissolution. In fact, there is a large positive, but not significant, association between shared eating time with television watching and the likelihood of relationship dissolution.