Background Previous studies of students' chemistry achievement have indicated that students have difficulty in solving quantitative chemistry problems (Selvarantnam,1974,1983; Powell,1977; McCabe,1977; Krammers-Pals et al., 1982, 1983; Gabel and Sherwood, 1983; Gabel et al., 1984; Genyea, 1983; Reif and Heller, 1982; Gilbert, 1980). Gil Perez (1982), Gil Perez and Terrogrosa (1981, 1983), Luria (1976) and Larkin (1980) observed that problem-solving was very difficult for secondary school students, and one of the principal causes of failure in school science and mathematics, because it is a complex intellectual task. Empirical studies in Nigeria, also, indicate the same results (Bello,1985; Adeyegbe,1985; Bello and Bajah,1987; Egbugara, 1989; Adigwe, 1991, 1992b, 1993). Research reports generally indicated that students' difficulties are associated with lack of procedural knowledge/strategies and skills of solving problems, and the reasoning skills that go along with them rather than lack of the conceptual knowledge and skills of chemistry (Gil Perez and Terrogrosa, 1981, 1983; Adeyegbe, 1985; Ahmed, 1982; Frazer and Sleet, (1984) Adigwe, (1993).
The traditional approach to teaching chemistry problem-solving involves the representation of worked examples in textbooks. Most worked examples do not teach the effective processes of solving chemistry problems (Selvarantnam, 1974, 1983; Frazer, 1982; Bello and Bajah, 1987). The traditional approach does not therefore teach the basic procedural knowledge/strategies and skills of solving quantitative problems. The implication is that the students do not acquire the problem-solving procedures and skills required for successful performance (Thorsland and Novak, 1974; Powell, 1977; Stewart,1982; Kempa and Nicholls, (1983); Selvarantnam and Frazer, 1982; Metres et al., 1980, 1981; McCabe, 1977). Literature review indicates that different instructional strategies could help to improve students' problem-solving skills (Chippetta and Russell, 1982; Payne, 1984; Decorte and Somers, 1982; Bessener and Smith, 1972; Polya, 1957; Butts and Jones, 1966; Newell and Simon, 1972; Mettes et al., 1980, 1981; Reif et al., 1976; Reif, 1983). It was therefore important for appropriate instructional strategies that could help students to understand the procedures/strategies and acquire the basic skills of solving quantitative chemistry problems to be developed. Efforts at developing instructional strategies to enhance students' problem-solving skills in chemistry led to the development of the following models: Selverantnam and Frazer (1982); the programme of Action and Methods (PAM) of Mettes et al., (1980, 1981); and the network approach, involving links and nodes (Ashmore et al., 1979; Elliot, 1982). The effects of the PAM and Selvarantnam and Frazer (1982) model on students' problem-solving achievement have been investigated: Mettes et al. (1980, 1981) observed that undergraduate students' skills in solving thermodynamics problems improved significantly if the four-stage model (PAM) they developed was coupled with mastery learning strategy. Bello (1985) found that the Selvarantnam and Frazer (1982) model significantly improved secondary school students' skills in solving stoichiometric problems if coupled with either practice, verbal feedback and remedial instructions or with practice and verbal feedback. Bello's (1985) study has remained the only research attempt to investigate the effects of a problem-solving instructional strategy on Nigerian students' problem-solving ability in chemistry. There is therefore a need to investigate empirically the relative effects of other instructional strategies on Nigerian chemistry students attainment in problem-solving.
Since Polya's (1957) heuristic approach to solving mathematics problems is becoming popular among Nigerian mathematics teachers (Agwagah, 1993) and PAM is gradually appearing in Nigerian science/chemistry literature as a problem-solving instructional approach (Egbugara, 1989; Bello, 1985; Bellow and Bajah, 1987, it was considered necessary to investigate empirically the relative effects of Polya's (1957) model, Mettes et al. …